JP2024012879A - Method for producing tumor-reactive t-cell - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique for minimizing time, cost, and labor, and which is capable of efficiently selecting a tumor-reactive T-cell simply by using an expression molecule marker as an indicator.

SOLUTION: A method for producing a tumor-reactive T-cell comprises: (1) a candidate isolation process of isolating a CD-8 positive T-cell from a tumor-infiltrating T-cell (Tumor-Infiltrating Lymphocyte: TIL) or whole blood collected from a cancer patient by using PD-1 and/or 4-1BB of an antigen of a cell surface as a marker; (2) a tumor-reactive evaluation process of transducing a T-cell receptor (TCR) gene of a T-cell isolated in the candidate isolation process into the T-cell, and specifying/evaluating a TCR equipped with tumor-reactivity; and (3) a therapeutic T-cell acquisition process of propagating a T-cell reacting a tumor in vitro to obtain a tumor therapeutic T-cell to be administered to the cancer patient.

SELECTED DRAWING: Figure 1

COPYRIGHT: (C)2024,JPO&INPIT

Description

The present invention relates to methods for producing tumor-reactive T cells, etc.

In recent years, advances in anti-checkpoint antibody therapy that use antibodies against molecules such as PD-1 and PD-L1 have enabled cancer patients to recognize their tumors, especially mutated antigens caused by genomic mutations in individual tumors. It is clear that there are T cells that recognize this, and that reactivation of suppressed T cells can lead to tumor regression and prolongation of prognosis.
In addition, "TIL therapy" in which expanded cultured TILs (Tumor-infiltrating T cells) are infused has been shown to be effective, especially for malignant melanoma. Specific T cells have been reported.
Furthermore, in recent years, the administration of T cells newly endowed with tumor reactivity by chimeric antigen receptors (CARs) or T cell receptors (TCRs) (CAR-T cells, TCR-T cells) has been shown to have tumor regression effects. It is also clear that this can be expected.
Therefore, there are growing expectations that treatments using tumor-reactive T cells that recognize tumor antigens, including mutated antigens, or their TCRs will be effective treatments for solid cancers. Against this background, attempts have been made to identify and obtain tumor-reactive T cells in various cancers, but research has mainly focused on malignant melanoma, and at present, research has focused on malignant melanoma and other cancers including colorectal cancer. In reality, it has been extremely difficult to reliably obtain tumor-reactive T cells not only from peripheral blood but also from TIL, where many tumor-reactive T cells are expected to exist in cancer (Non-Patent Documents 1 to 3). ).

Publication WO2014-017533A1

PD-1 identifies the patient-specific CD8+ tumor-reactive repertoire infiltrating human tumors. J Clin Invest. 2014 May;124(5):2246-59. Bystander CD8+ T cells are abundant and phenotypically distinct in human tumor infiltrates. Nature. 2018 May;557(7706):575-579. Isolation of T-Cell Receptors Specifically Reactive with Mutated Tumor-Associated Antigens from Tumor-Infiltrating Lymphocytes Based on CD137 Expression. Clin Cancer Res. 2017 May 15; 23(10): 2491-2505. Fujii et al., Identification of an immunogenic neo-epitope encoded by mouse sarcoma using CXCR3 ligand mRNAs as sensors, Oncoimmunology. 2017 Mar 20;6(5):e1306617. Hanson HL.et al., Eradication of Established Tumors by DUC18T Cell Adoptive Immunotherapy, Immunity 2000, 13:265-76.

It has been reported that the expression of PD-1, LAG3, and TIM3 in TILs can be an indicator of tumor-reactive T cells, but this study was conducted in malignant melanoma, and appropriate treatment is not recommended for other cancers. No indicators were found.
In addition, there are reports that (1) it is possible to obtain tumor-specific T cells by co-culturing them with TILs and selectively culturing and proliferating tumor-specific T cells, such as antigen-presenting cells that have taken up peptides or tumor lysates; 2) A method has been reported in which single-cell RNA sequencing and TCR sequencing are performed on TILs and tumor-reactive T cells are selected using a characteristic group of molecules expressed at the mRNA level as indicators. However, these methods are not practical because they require a lot of time, effort, and cost when applied to individual cancer patients.
The present invention was made in view of the above-mentioned circumstances, and its purpose is to minimize time, cost, and labor, and to easily and efficiently select tumor-reactive T cells using expressed molecular markers as indicators. The goal is to provide the technology that makes it possible.

The method for producing tumor-reactive T cells according to the invention for solving the above problems is as follows: (1) From tumor-infiltrating T cells (Tumor-infiltrating lymphocytes: TILs) or whole blood collected from cancer patients, a candidate isolation step of isolating CD-8 positive T cells using PD-1 and/or 4-1BB as a marker; (2) a T cell receptor of the T cells isolated in the candidate isolation step; (TCR) gene transfection into target cells to identify and evaluate TCRs with tumor reactivity; (3) T cells with tumor-reactive TCRs are expanded in vitro; The present invention is characterized by comprising a therapeutic T cell obtaining step for obtaining tumor therapeutic T cells to be administered to cancer patients.
In the above invention, target cells refer to cells that have the ability to exhibit tumor reactivity by expressing a predetermined TCR gene, such as T cells and peripheral blood mononuclear cells (peripheral blood mononuclear cells). mononuclear cells (PBMC).
In the above invention, the solid tumor is brain tumor, tongue cancer, esophageal cancer, stomach cancer, small intestine cancer, colon cancer, liver cancer, kidney cancer, bladder cancer, lung cancer, thyroid cancer, breast cancer, uterus cancer, Preferably, the cancer is at least one selected from the group consisting of cancer, ovarian cancer, prostate cancer, rhabdomyosarcoma, and leiomyoma.
Moreover, it is preferable that the candidate isolation step further includes using tetraspanin as a marker. At this time, it is preferable that the tetraspanin is at least one selected from the group consisting of CD9, TSPAN2, CD151, CD53, CD37, CD82, CD81, and CD63.

<Summary of the present invention>
1. Method for identifying TCRs that respond to autologous tumors Figure 1 outlines a method for searching for cell surface antigens of tumor-reactive T cells. Tumor tissue is obtained from cancer patients with solid tumors (for example, colon cancer patients) (1). Generally, tumor tissue contains tumor cells and TILs in addition to normal cells. Patient tumor cells are maintained in vitro by creating cancer tissue-originated spheroids (CTOS) from the tumor cells contained in the tumor tissue (5). On the other hand, CD8 ⁺ TILs were identified from tumor tissues (2), and CD8 ⁺ T cell populations were subdivided using PD-1 and 4-1BB expression as indicators, and each cell population (PD- 1 positive and 4-1BB negative (PD-1 ⁺ 4-1BB ⁻ ), PD-1 positive and 4-1BB positive (PD-1 ⁺ 4-1BB ⁺ ), both PD-1 and 4-1BB negative (PD- 1 ^- 4-1BB ^- ))) (candidate isolation step), obtain the T cell receptors (TCRs) of the T cells by single-cell PCR, and determine the gene sequences of these TCRs. Perform repertoire analysis using this method (3). Individual TCRs, mainly TCRs with large repertoire sizes, were expressed and introduced into peripheral blood mononuclear cell (PBMC) fractions of healthy individuals using retroviral vectors (4), and autologous tumor-derived CTOS and 24 By co-culturing for a period of time and measuring the cytokine (IFN-γ) in the culture supernatant, it is possible to identify TCRs that react with autologous tumors (6: Tumor reactivity evaluation step).
The upper right corner of the figure shows an image obtained when CEA (carcinoembryonic antigen), which is highly expressed in colorectal cancer tissue, was stained with an anti-CEA antibody to confirm the "tumor nature" of CTOS. .

2. Method for verifying the ability of TCR to recognize mutant antigens Figure 2 shows a verification method for analyzing the reactivity of tumor-reactive TCRs that react with CTOS to mutant antigens (antigens derived from genetic mutations in individual tumor genomes). showed that.
Nucleic acids (DNA/RNA) obtained from individual patient tumor tissues (1) and normal tissue nucleic acids (DNA) obtained from patient peripheral blood (2) are analyzed by next-generation sequencing (3), and tumor-specific Identify genetic mutations that involve amino acid mutations (4). Furthermore, in silico analysis is performed to select mutant antigens that are likely to be presented to the patient's HLA and are expected to be highly immunogenic (5). We construct a plasmid (Tandem Mini Gene; TMG) in which 8 to 10 base sequences encoding these mutated antigens (27-mer amino acid sequence (mutated amino acid in the center, 13 amino acids on both sides)) are arranged in tandem ( 6), further produce mRNA by in vitro transcription.
On the other hand, CD8 ⁺ TILs are identified from patient tumor tissues and the clonality of the isolated CD8 ⁺ TILs is analyzed (7). Transduce the TCR gene (8) and evaluate tumor reactivity using CTOS (9: candidate isolation step).
Expression of the previously prepared mRNA is introduced into LCL (immortalized B cell line) prepared from patient B cells using electroporation, and co-cultured with T cells into which tumor-reactive TCR has been introduced. and analyze its reactivity using the ELISPOT method (tumor reactivity evaluation step). When TCR recognizes a specific mutated antigen, it reacts with LCL into which TMG mRNA has been introduced. In this case, since there are 8 to 10 types of mutated antigens in TMG, it is possible to determine which mutated antigen is recognized by creating new individual TMGs and analyzing them (10).

3. Immune cell therapy using TIL or PBMC from cancer patients Figure 3 shows an overview of immune cell therapy using TIL or PBMC derived from cancer patients.
Tumors or PBMCs are collected from cancer patients, and PD-1 ⁺ 4-1BB ⁺ CD9 ⁺ triple positive CD8 ⁺ T cells are obtained as TIL candidates (1: candidate isolation step). From here, the TCR gene is analyzed and a tumor-reactive TCR is obtained (2). Next, this TCR is transduced into T cells (3: tumor reactivity evaluation step). Separately, PD-1 ⁺ 4-1BB ⁺ CD9 ⁺ tumor-reactive TCRs can be obtained from patient PBMCs (4). Next, the cells obtained in (3) and/or (4) are grown in vitro (5: therapeutic T cell acquisition step), and then administered to the patient (6). By administering these T cells, it is possible to induce an immune response that changes a tumor (cold tumor with little immunogenicity) to a tumor (hot tumor) with strong immunogenicity (7), leading to a cure. I can (8).

According to the present invention, T cells for tumor treatment of cancer patients can be provided. These T cells can be effectively used to treat cancer patients. Examples of such treatment methods include the following:
1. Genetically modified T cell infusion therapy using TCR of tumor-reactive T cells, 2. 3. A treatment method in which a population of tumor-reactive T cells is expanded and cultured without modification and then infused; 3. 4. A treatment method that increases efficacy by co-expressing tetraspanin molecules simultaneously with genetic modification with TCR/CAR; and 4. This includes the application of extracellular vesicles (EVs) containing exosomes from tumor-reactive T cells to cancer treatment.

FIG. 2 is a schematic diagram illustrating a method for searching for cell surface antigens of tumor-reactive T cells. FIG. 2 is a schematic diagram illustrating a verification method for analyzing reactivity to mutant antigens for tumor-reactive TCRs that react with CTOS. FIG. 2 is a schematic diagram of immune cell therapy using TIL or PBMC derived from cancer patients. FIG. 2 is a diagram showing the results of analyzing tumor-reactive TCR for CCP1. (A) Pie chart showing the results of dividing TIL into three types of cell populations depending on the presence or absence of 4-1BB and PD-1 expression, (B) Whether DP-1 to DP-9 produce IFN-γ (C) A graph showing the results of analysis using the ELISPOT method after introducing mRNA prepared from four types of TMG into CCP1 patient-derived LCL. FIG. 3 is a diagram showing the results of analyzing tumor-reactive TCR for CCP3. (A) Pie chart showing the results of dividing TIL into three types of cell populations depending on the presence or absence of 4-1BB and PD-1 expression, (B), (C) Using four types of TMG, CCP3 patient-derived FIG. 2 is a diagram showing the results of examining reactivity after expression was introduced into LCL. FIG. 2 is a diagram showing the results of analyzing tumor-reactive TCR for CCP5. (A) A pie chart showing the division of TIL into two types of cell populations based on the presence or absence of PD-1 expression, and a diagram showing the results of repertoire analysis obtained by acquiring individual TCR genes from each cell population, (B) Graph showing the results of analysis by ELISPOT after expression of mRNA prepared from two types of TMG was introduced into LCL derived from a CCP5 patient. (C) Results of investigating the reactivity of TMs in which genes in TMG1 were individually divided. This is a graph showing. FIG. 3 is a diagram showing the results of analyzing tumor-reactive TCR for CCP15. (A) Pie chart in which TILs are divided into three types of cell populations depending on the presence or absence of PD-1 and 4-1BB expression. (B) After expression of mRNA prepared from four types of TMG is introduced into LCLs derived from CCP15 patients. , a graph showing the results of analysis using the ELISPOT method, and (C) a graph showing the results of examining the reactivity of TMs in which genes in TMG1 were individually divided. FIG. 3 is a diagram showing the results of investigating the expression frequency of PD-1 and 4-1BB among tumor-reactive TCRs among CD8 ⁺ TILs obtained from 12 cases of CCP. (A) Diagram comparing the percentage of DP (Double Positive) population of PD-1 ⁺ 4-1BB ⁺ (B) Diagram showing the results of examining the presence or absence of expression of PD-1 and 4-1BB for CCP1, (C) For CCP1, this is a pie chart in which TILs are divided into three types of cell populations based on the presence or absence of PD-1 and 4-1BB expression. This is a graph showing the results of a log-rank test regarding the prognosis of 16 cases based on "PD-1 ⁺ 4-1BB ⁺ cell population frequency in CD8 ⁺ TIL of 3% or more". FIG. 3 is a diagram showing the results of sc (single cell) RNA-seq analysis and TCR-seq analysis using CCP1 case TIL. (A) Pie chart showing the results of investigating the PD-1 ⁺ 4-1BB ⁺ DP population and graph showing the results of investigating the reactivity to CTOS, (B) Approximate RNA expression profiles for the two types of TCR (C) A diagram showing the results of RNA expression analysis of DP5-6 and DP53-1 and other cells. FIG. 2 is a diagram showing the results of analyzing tumor-reactive TCR for CCP9. (A) Pie chart showing the results of selecting by PD-1 ⁺ and examining TCR, (B) Selecting by TP (Triple Positive) of PD-1 ⁺ 4-1BB ⁺ CD9 ⁺ and examining the results of TCR. (C) A graph showing the results of investigating whether SP12-2 and the like produce IFN-γ. FIG. 3 is a diagram showing the results of examining whether the expression of CD9 molecules can serve as an indicator of tumor reactivity using a mouse model. (A) Diagram showing the collection schedule of CD8 ⁺ T cells used in the IFN-γICS method, (B) Graph showing the results of implementing the IFN-γICS method, (C) Mutant Snd1 peptide and adjuvant (CHP: hydrophobized This figure shows the results of an investigation into the effect of polysaccharides) on CMS7 tumor-bearing mice (the upper part is the experimental protocol, and the lower part is a graph examining the relationship between tumor length and tumor administration). FIG. 2 is a diagram showing the results of examining whether mutant Snd1-specific CD8 + T cells express CD9 molecules. (A) Diagram showing the collection schedule of CD8 + T cells, (B) Graph showing the results of implementing the IFN-γICS method, (C) The frequency of mutant Snd1 peptide-specific CD8 ⁺ cells was investigated by the ELISPOT method. It is a graph showing the results. FIG. 3 is a diagram showing the results of TCR repertoire analysis of CD8 ⁺ CD9 ⁺ CD81 ⁺ T cells. (A) Diagram showing the collection schedule of splenocytes, (B) Diagram showing the results of TCR analysis of CD8 ⁺ CD9 ⁺ CD81 ⁺ T cells in the spleen on Day 21, (C) CD8 ⁺ specific for mutant Snd1 FIG. 2 is a diagram showing the results of analyzing the expression of CD81 in T cells. FIG. 3 is a diagram showing the results of TCR repertoire analysis of CD8 ⁺ CD9 ⁺ CD81 ⁺ T cell population in the spleen of ICI-untreated mice after CMS7 tumor bearing and analysis of TCR reactivity to CMS7. (A) Graph showing the measurement results of mouse IFN-γ (mIFN-γ), (B) Graph showing the measurement results of mouse TNFα (mTNFα), (C) TCR that specifically recognizes CMS7 was observed in 3 cells. This is a pie chart showing the results obtained. FIG. 3 shows the results of TCR repertoire analysis of CD8 ⁺ CD9 ⁺ CD81 ⁺ T cell populations in the spleen of CT26 tumor-bearing mice. (A) A diagram showing the collection schedule of mouse spleen, and (B) a diagram showing the results of repertoire analysis of CD9 ⁺ CD81 ⁺ cells among CD8-positive cells in the spleen on Day 21. FIG. 3 is a diagram showing the results of examining the characteristics of CD8 ⁺ CD9 ⁺ CD81 ⁺ T cell populations in the spleen of CT26 tumor-bearing ICI-untreated mice. (A) A diagram showing the schedule for collecting mouse spleen, (B) and (C) graphs showing the measurement results of mouse IFN-γ (mIFN-γ). FIG. 3 is a diagram showing the results of an experiment to verify that CD9 expression enhances antitumor effects. (A) A diagram showing the test schedule, (B) a graph showing the relationship between tumor administration and tumor size.

Next, embodiments of the present invention will be described with reference to diagrams, but the technical scope of the present invention is not limited by these embodiments, and various forms may be implemented without changing the gist of the invention. It can be carried out with
<Materials and test methods>
1. Ethics regarding human samples Written informed consent was obtained from patients and healthy volunteers in accordance with the guidelines of the Declaration of Helsinki. The study protocol was approved by the Ethics Committee of Mie University School of Medicine.
2. Ethics regarding mouse experiments All animal experiments were performed using protocols approved by the Mie University Life Science Center Animal Care and Use Committee.
3. Mice Female BALB/c mice were purchased from Shizuoka Animal Research Institute. α/β-TCR transgenic DUC18 mice that react with H-2K ^d mERK2 _136-144 were established as previously reported. NOD/Shi-scid/IL-2Rγ ^null mice, known as NOG mice, were purchased from the Central Institute of Laboratory Animals. All mice were housed in a specific pathogen-free environment and used at 8-10 weeks of age.

4. Antibodies Fluorescein isothiocyanate (FITC)-conjugated anti-human CD3 monoclonal antibody (mAb, OKT3), phycoerythrin (PE)-Cy7-conjugated anti-human CD8 mAb (RPA-T8), PerCP-Cy5.5-conjugated anti-human CD9 (HI9a) , allophycocyanin (APC)-conjugated anti-human PD-1 mAb (EH12.2H7), PE-conjugated anti-human 4-1BB (CD137) mAb (4B4-1) and eFluor780-Fixable Viability Dye were purchased from BioLegend. PerCP-Cy5.5-conjugated anti-mouse CD3 mAb (145-2C11), APC-Cy7-conjugated anti-mouse CD8 mAb (53-6.7), FITC-conjugated anti-mouse CD9 mAb (MZ3), PE-conjugated anti-mouse CD81 mAb (Eat-2 ) and PE-Cy7-conjugated anti-mouse PD-1 mAb were purchased from BioLegend. FITC-conjugated anti-CEA mAb (CB30) was purchased from Abcam. For the treatment of mice, anti-mouse PD-1 (RMP1-14), anti-mouse CTLA-4 (9D9) and anti-mouse GITR (DTA-1) mAbs were produced from hybridomas and purified on a protein G column.
5. cell line
CMS5 and CMS7 are 3-methylcholanthrene-induced sarcoma cell lines derived from BALB/c. CT26 is a colon epithelial tumor cell line derived from intrarectal injection of N-nitroso-N-methylurethane into BALB/c mice. P1.HTR is a subline of the P815 mastocytoma cell line derived from DBA/2. Human B-lymphoblastoid cell lines (LCL) were generated in our laboratory from patients' peripheral blood using EBV-containing supernatants. All mouse tumor lines and LCLs were cultured in RPMI-1640 medium supplemented with 10% heat-inactivated fetal calf serum (FCS), 50 μM 2-mercaptoethanol (2-ME), and 0.2 mg/mL glutamine.

6. Patients and Samples Patients with colorectal cancer (CRC) who visited the Department of Gastrointestinal and Pediatric Surgery at Mie University Hospital and underwent surgery between 2016 and 2018 were enrolled in this study. Fresh tumor tissue samples were collected from surgically resected CRCs. Paired peripheral blood mononuclear cells (PBMCs) from individual patients were isolated from fresh heparinized venous blood samples by density gradient centrifugation on Ficoll-Paque PLUS (GE Healthcare). .
7. Preparation of Cancer Tissue-Originated Spheroid (CTOS) lines and TILs
CTOS was prepared by making appropriate changes to previous reports. A brief explanation is as follows. The excised tumor was immediately mechanically cut into small pieces and incubated at 37°C in DMEM/Ham F12 medium (manufactured by Wako Pure Chemical Industries, Ltd.) containing Liberase DH (manufactured by Roche) at a final concentration of 26 units/mL. Incubation was continued for 90 minutes with continuous stirring. After adding 10 μg/mL of DNase I (manufactured by Roche), the cells were cultured for an additional 15 minutes. The digested culture fluid was sequentially filtered using 500, 250, 100 and 40 μm mesh filters (BD Falcon). Fractions between 100 and 250 μm (Fr. 100-250) and 40 and 100 μm (Fr. 40-100) were collected. The Fr.40-100 sample was cultured using STEM PRO hESC SFM (manufactured by Invitrogen) at 37°C in an environment of 5% CO ₂ and 20% O ₂ for 24 to 48 hours to form CTOS. The Fr. 100-250 samples were mechanically disrupted using a 27-gauge needle by moving the plunger up and down several times, and then cultured similarly to the Fr. 40-100 samples.
Cells that had passed through a 40 μm mesh filter were collected, frozen at −80° C. using Cellbanker-1 medium (manufactured by Takara Bio Inc.), and used as TIL.

8. Purification of DNA and RNA from human samples A 5 mm diameter tumor tissue sample was used to isolate DNA and RNA. Purification of RNA and genomic DNA was performed using the RNeasy and Puregene Core kit (Qiagen) according to the manufacturer's protocol. 5×10 ⁶ PBMCs were used for DNA isolation. Purification of genomic DNA from PBMCs was carried out using the QIAmp DNA kit (manufactured by Qiagen) according to the manufacturer's protocol.
9. Next-generation sequencing and data analysis Genomic DNA extracted from patient normal tissue and tumor tissue was cut into 150 to 200 bp using 200 ng, and after ligating adapters using SureSelect XT HS kit (manufactured by Agilent technology), the Exon region was extracted. Hybridization was performed using a biotinylated oligo RNA bait with an MHC region added to SureSelect Human All Exon V6. After collecting samples using avidin beads and performing PCR amplification to obtain a library, 101 bp paired-end Whole Exome Sequence (WES) was performed using NextSeq550 (manufactured by Illumina). Using 300-500 ng of total RNA extracted from tumor tissue, ligate an adapter to the mRNA using the TruSeq ^TM Stranded mRNA Library kit (manufactured by Llumina), perform PCR amplification to obtain a library, and then use NextSeq550. 76 bp paired-end RNA sequencing (RNA-Seq) was performed.

10. Identification of somatic mutations in tumor tissue We used the human reference genome hg38 to perform a comparative analysis of WES data of normal and tumor tissues from patients, and detected tumor-specific somatic mutations. Germline mutations were also detected using the same reference genome and WES data of normal tissues. Each analysis was performed in accordance with GATK best practices. After filtering the detected mutations to narrow them down to those with high accuracy, we used the GENCODE human reference gene model to extract mutations located on the protein translation sequence of the gene. Physical phasing was performed for somatic mutations and germline mutations that exist on the protein translation sequence of the same gene and are close to each other in terms of genomic locus, and the haplotypes surrounding the alleles with somatic mutations were determined. Furthermore, based on this haplotype information, we excised a total pattern of peptide sequences of 8 to 11 residues that could be the mutant antigen epitope.
In order to quantify the transcriptional expression level of mRNA that translates the mutant antigen, we performed transcriptional expression analysis using RNA-Seq data of patient tumor tissues and quantified it at the isoform level using TPM (Transcript Per Million). Ta. A series of analyzes were performed using a method similar to the GTEx/TOPMed RNA-Seq pipeline. When there are multiple transcript isoforms that overlap with the same somatic mutation, the respective TPM values were summed to form TPM_SUM. Furthermore, in order to estimate the allele-specific transcription expression level including somatic mutations, we calculated the mutation allele frequency on the RNA Seq data of the somatic mutations detected by WES, and multiplied this by the TPM or TPM_SUM value. Calculated as TPMvar or TPM_SUMvar.

In order to rank the possibility of presentation of peptide sequences of 8 to 11 residues, which can be variant antigen epitopes, to the human major histocompatibility complex (MHC), a score was calculated using a presentation prediction model developed by BrightPath Bio. Ta. The proposed prediction model was constructed using a generalized linear model (logistic regression model) using immunopeptidome data and random peptides obtained from public databases as learning data, and was constructed using the MHC binding prediction tools NetMHCpan-4.0 and MHCflurry-1.4, as well as proteasome cleavage. Parameter estimation was performed using the maximum likelihood method using the predicted values of the prediction tool NetChop-3.1 as explanatory variables. Using a linear predictor with these estimated parameters, the likelihood of presenting a patient-derived variant antigen epitope was scored for each HLA allele possessed by the patient. Specifically, as shown in Equation (1), the calculated value (z) of the linear predictor for each peptide was subjected to soft plus conversion and was used as the presentation score (SCORE). Furthermore, SCOREadj was calculated using equation (2) to perform correction based on the transcriptional expression level.

The HLA allele with the highest SCORE value among the HLA alleles possessed by the patient was used as the predicted HLA allele for the mutated antigen epitope, and the mutated antigen epitopes were ranked in descending order of the HLA allele's SCOREadj.

11. Preparation and in vitro transcription of DNA encoding tandem minigene (TMG)
The DNA sequence used in the TMG method was designed as follows. A 27mer peptide sequence centered around the mutant residues was designed and back translated into a DNA sequence. In the case of insertion/deletion somatic mutations, back translation into DNA sequences was performed using a 27-mer peptide sequence centered on the 8- to 11-mer mutant antigen epitope. TMG was cloned using BamHI and XhoI in the MCS of pcDNA3.1(+). After linearizing the plasmid DNA by cutting with the XhoI enzyme, the DNA was precipitated using sodium acetate and ethanol. Next, 1 μg of DNA was used to generate in vitro transcribed RNA using the mMESSAGE mMACHINE T7 Ultra kit (manufactured by Life Technologies) according to the manufacturer's protocol. The resulting capped and tailed RNA was resuspended in water at 1 μg/μL and stored at −80° C. before use for LCL transfection.

12. electroporation
LCLs were harvested, washed twice with phosphate buffered saline (PBS), and resuspended in RPMI1640 medium. Next, 10 μg of mRNA was mixed with 100 μl of cell suspension containing up to 1 × ¹⁰ cells, transferred to a 2 mm gap cuvette (Genetronics), and transferred to a BTXECM830 square wave electroporator (Genetronics). ) was used for electroporation. After electroporation, cells were immediately transferred to 2.0 ml of medium and cultured overnight in a 24-well plate at 37 °C in a CO ₂ incubator until use.
13. Intracellular staining (ICS)
Intracellular cytokine staining was performed as follows. Splenocytes (1×10 ⁶ ) were incubated with 50 μM synthetic peptide or DMSO for 15 minutes at room temperature, followed by GolgiPlug (BD Bioscience) for 4 hours. Cells were stained with CD8α-specific mAb, CD9-specific mAb, or CD81-specific mAb for 15 min at 4 °C. Cells were stained with allophycocyanin-conjugated anti-IFNγ mAb after permeabilization and fixation using the Cytofix/Cytoperm kit (BD Biosciences) according to the manufacturer's protocol.

14. Enzyme-linked immunosorbent assay (ELISA)
Human or mouse cytokine concentrations in culture supernatants were determined using an ELISA kit (Thermo Fisher Scientific) according to the manufacturer's protocol.
15. Enzyme-linked immunospot (ELISPOT) assay The human IFNγ ELISPOT assay was performed as previously described with minor modifications. Briefly, ELISPOT plates (MAHA S4510, Millipore) were coated with anti-human IFNγ mAb (1-D1K, Mabtech). A total of 2 × 10 ⁴ TCR-transduced T cells and 2 × 10 ⁴ TMG mRNA-transduced LCLs were added to each well of the plate. After incubation at 37°C for 22 hours, the plate was washed, biotinylated capture antibody (7-B6-1, manufactured by Mabtech) was added, and the plate was incubated at 4°C overnight. After washing the wells, the cells were reacted with alkaline phosphatase-labeled streptavidin, and then stained with an alkaline phosphatase-labeled substrate kit (manufactured by Bio-Rad). Spots were counted using an ELISPOT plate reader (ImmunoSpot, manufactured by CTL-Europe GmbH).

16. peptide
MuLV gp70-derived AH-1 (SPSYVYHQF (SEQ ID NO: 1)), mutated ERK2-derived peptide (mERK2-9m: QYIHSANVL (SEQ ID NO: 2)), mutated Snd1-derived peptide (mSnd1: YAPCRGEF (SEQ ID NO: 3)), and non-mutated The Snd1-derived peptide (wSnd1: YAPRRGEF (SEQ ID NO: 4)) was obtained from Invitrogen with a purity of over 80%. All peptides were dissolved in DMSO at a concentration of 10 mM, aliquoted and stored at -80°C before use.
17. Performing isolated cell RNA sequencing and TCR sequencing
CCP1 TILs were washed multiple times and then resuspended in 1× PBS containing 0.04% BSA. Approximately 5 x ¹⁰ cells with 100% viability were used for generation of GEMs (gel beads in emulsion) using Chromium Controller (10x Genomics). By this method, RNA molecules were tagged individually in each cell. After converting molecularly tagged RNA fragments containing those derived from the TCR V(D)J gene into cDNA, Chromium Single Cell V(D)J Reagent Kits v1.1 (manufactured by 10× Genomics) were used. , established a library for next-generation sequencing. The obtained library was quantified using Qubit dsDNA Assay (manufactured by ThermoFisher Scientific) and TapeStation D1000 (manufactured by Agilent). Gene expression libraries and V(D)J libraries were each pooled at a ratio of 9:1 and sequenced using the Illumina HiSeq platform with the following configuration. That is, read 1: 26 bp, read 2: 91 bp, and i7 index: 8 bp, for a total of about 400 M paired-end reads per sample. Image analysis and base calling were performed by HiSeq instrument software. FASTQ read raw data were imported into the Cell Ranger V(D)J pipeline (10x Genomics) for mapping, gene expression counting, and clonotype annotation. The above isolated cell analysis was performed according to the manufacturer's protocol.

18. Preparation of PBMC
PBMCs were separated from heparin-treated fresh venous blood samples by density gradient centrifugation using Ficoll-Paque PLUS (manufactured by GE Healthcare).
19. Preparation of Mouse Splenocytes Mouse spleen was crushed using a microscope slide, treated with ammonium potassium chloride (ACK), and filtered through Cell Trics® 30 μm (manufactured by Sysmex Partec). Isolated cell suspensions from pooled spleens were cultured with appropriate peptides in 96-well V-bottom plates at a density of 5×10 ⁶ cells/ml. In some experiments, CD8 ⁺ splenic T cells were obtained by positive enrichment using the MACS system (Miltenyi Biotec). Flow cytometry confirmed that the T cell fraction contained >95% CD8 ⁺ T cells.
20. Flow cytometry analysis and isolated cell sorting
TILs were thawed and stained with FITC-anti-human CD3 mAb, PE-Cy7 anti-human CD8 mAb, PE-anti-human CD137(4-1BB) mAb and APC anti-human PD-1 mAb. Dead cells were excluded by staining with eFluor780-Fixable Viability Dye. After staining, cells were analyzed using FACSAria (Beckton Dickinson). CD8 ⁺ PD-1 ⁺ CD137 ⁺ , CD8 ⁺ PD-1 ⁺ CD137 ^- and CD8 ⁺ PD-1-CD137 ^- T cells were isolated into 96-well PCR plates, respectively. For mice, splenocytes were stained with PerCP-Cy5.5-conjugated anti-mouse CD3 mAb, APC-Cy7-conjugated anti-mouse CD8 mAb, FITC-conjugated anti-mouse CD9 mAb, and PE-conjugated anti-mouse CD81 mAb. Dead cells were excluded from staining with eFluor780-FixableViabilityDye.

21. Amplification of TCR cDNA from isolated T cells
Cloning of T cell receptor (TCR) followed the method described in WO2014-017533A1 (Patent Document 1). A brief explanation is as follows. RNA extracted from each isolated T cell was converted to cDNA, and Vα and Vβ regions were amplified by multistep nested PCR. Thereafter, the DNA sequence of the PCR product was determined by direct sequencing. The TCR repertoire was analyzed using the IMGT/V-Quest tool (http://www.imgt.org/). For transduction of human PBMCs or mouse splenocytes, TCRα and TCRβ chains were linked by the P2A sequence and cloned into the pMX-IRES-EGFP vector.
22. Production of retroviruses
Plat-A or Plat-E cells (Cell Biolabs) were used to generate retroviruses for transducing TCR genes into human PBMCs or mouse splenocytes.
23. Retroviral transduction of human PBMCs PBMCs obtained from healthy volunteers were incubated with anti-CD3 antibody-coated plates and RetroNectin (Takara Bio Inc.) in the presence of 600 IU/mL recombinant IL-2. Used to stimulate. Proliferating lymphocytes were transduced with retroviruses encoding candidate TCRα/β genes and expanded in vitro. Ten days after transduction, cells were harvested and used for experiments.

24. Retroviral transduction of mouse splenocytes
Whole splenocytes (1.5 × ¹⁰ cells/5 mL) from BALB/c mice were incubated with immobilized anti-CD3 (1 μg/mL; 145-2C11) and soluble anti-CD28 (1 μg/mL) in one well of a 6-well plate. ; 37.51). One day after stimulation (day 1), ^each well of a 24-well plate was coated with RetroNectin (manufactured by Takara Bio Inc.) for 5 × 10 cells, and after adding 1 mL of medium, the RetroNectin-conjugated virus infection method was used. and transduced with a candidate TCR-expressing viral vector. On day 3, the cells were transferred to a 50 mL flask containing 10 mL of medium, and were grown and cultured. Cells were collected on day 5 and used for experiments. During the culture, 60 IU/mL of recombinant human IL-2 (manufactured by Novartis) was added. In some experiments, murine CD9-transduced T cells were prepared using splenic T cells obtained from DUC18 mice.
25. Single-cell RNA sequence analysis and TCR sequence analysis using TILs TILs that had been frozen and preserved in liquid nitrogen were thawed using 10% FBS RPMI1640 culture medium and washed several times using 2% FBS 1×PBS. Next, CD8 ⁺ T cells in TIL were separated using human CD8 magnet beads (manufactured by Miltenyi Biotec), and single-cell RNA-seq and TCR-seq were performed.
26. Statistical analysis Statistical analysis was performed by unpaired Student's t test, and less than 5% (p<0.05) was considered statistically significant.

<Test results>
Table 1 summarizes patient information, whether CTOS has been established, and whether the TCR gene that recognizes CTOS has been acquired.

As shown in the table, CTOS could be established in 10 of the 16 colorectal cancer patients (CCP1 to CCP16) (marked with "○" or "△" in the CTOS column). Furthermore, in 6 of these 10 cases, TIL-derived TCR could be obtained ("+" mark in the TCR CTOS-Reactivity column).
Figure 4 shows the results of analyzing tumor-reactive TCR for CCP1 among the six cases (CCP1, 3, 5, 7, 9, 15).
In CCP1 cases, CD8 ⁺ TILs are divided into three types of cell populations depending on the presence or absence of PD-1 and 4-1BB expression (in Figure 4 (A), the lower left (PD-1 ⁻ ,4-1BB ⁻ : 60 cells) ), lower right (PD-1 ⁺ ,4-1BB ^- : 72 pieces), upper right (PD-1 ⁺ ,4-1BB ⁺ : 72 pieces).In addition, PD-1 ^- ,4 corresponding to the upper left -1BB ⁺ was not detected.), individual TCR genes were obtained from each cell population, and repertoire analysis was performed. The pie chart shown in Figure 4(A) is the result of TCR repertoire analysis for each cell population, and the two corresponding TCRs are most frequently enriched in the PD-1 ⁺ 4-1BB ⁺ fraction (DP). It shows that
We also examined the reactivity of these TCRs to CTOS. As a result, as shown in FIG. 4(B), two types of TCR (DP5-6 and DP53-1; DP means Double Positive) showing tumor reactivity were obtained. Reactivity to CTOS was determined by the production of IFN-γ by co-culture of CTOS and TCR-transduced T cells.
Regarding the reactivity of these two types of TCR to mutated antigens, we created four types of TMG (TM1 to TM4), introduced the mRNA created from each TMG into CCP1 patient-derived LCL, and then analyzed it using the ELISPOT method. As shown in FIG. 4(C), both TCRs showed reactivity with TMG1. Further analysis revealed that both TCRs specifically recognized the mutant CREBBP in TMG1. The amino acid sequences in the figure are "NGTASQSTSPSQPCKKIFKPEELRQAL" (SEQ ID NO: 5) for mutant CRBBP and "NGTASQSTSPSQPRKKIFKPEELRQAL" (SEQ ID NO: 6) for wild type CRBBP.

Figure 5 shows the results of analyzing tumor-reactive TCR for CCP3.
In the CCP3 case, CD8 ⁺ TILs were divided into three types of cell populations based on the expression of PD-1 and 4-1BB (Figure 5(A)), and individual TCR genes were extracted from each cell population, as in Figure 4. were acquired and repertoire analysis was performed. Furthermore, as a result of examining the reactivity of these TCRs to CTOS, we obtained a TCR (DP72-6) that showed tumor reactivity (Figure 5(B)). The pie chart shown in FIG. 5(A) is the result of TCR repertoire analysis for each cell population. Two T cells expressing the relevant TCR were observed in the PD-1 ⁺ 4-1BB ⁺ DP fraction, and only one cell (single) in the PD-1 ⁺ 4-1BB ^- fraction. It shows.
Regarding the reactivity of this TCR to mutated antigens, we created four types of TMG (TM1 to TM4), introduced the mRNA created from each TMG into CCP3 patient-derived LCL, and analyzed it by ELISPOT, and found that it reacted with TMG1. (Figure 5(C)). Further analysis revealed that it specifically recognized the mutant DDX60 in TMG1. The amino acid sequences in the figure were "PRVMDMLKLYFLFYLQFLVKEGYLDQE" (SEQ ID NO: 7) for mutant DDX60 and "PRVMDMLKLYFLFSLQFLVKEGYLDQE" (SEQ ID NO: 8) for wild type DDX60P.

Figure 6 shows the results of analyzing tumor-reactive TCR for CCP5.
In the CCP5 case, CD8 ⁺ TILs were divided into two types of cell populations based on PD-1 expression (Figure 6(A)), individual TCR genes were obtained from each cell population, and repertoire analysis was performed. Furthermore, as a result of examining the reactivity of these TCRs to CTOS, we obtained a TCR (SP27-2) that shows tumor reactivity. The pie chart shown in Figure 6(A) is the result of TCR repertoire analysis in each cell population, and T cells expressing the relevant TCR were observed in 3 cells in the PD-1 ⁺ fraction, and PD-1 ^- Indicates that it was not observed in fractionation.
Regarding the reactivity of TCR to mutated antigens, we created two types of TMG (TM1 and TM2), introduced the mRNA created from each TMG into CCP5 patient-derived LCL, and analyzed it by ELISPOT, and found that it was reactive with TM1. (Figure 6(B)). Further analysis revealed that it specifically recognized mutant FAM129B in TM1 (Figure 6(C). 4 in TM1 is "FAM129B"). The amino acid sequences in the figure were "RAQIHMREQMDNAMYTFETLLHQELGK" (SEQ ID NO: 9) for mutant FAM129B and "RAQIHMREQMDNAVYTFETLLHQELGK" (SEQ ID NO: 10) for wild type FAM129B.

FIG. 7 shows the results of analyzing tumor-reactive TCR for CCP15.
In CCP15 cases, CD8 ⁺ TILs were divided into three types of cell populations based on the expression of PD-1 and 4-1BB (Figure 7(A)), and individual TCR genes were obtained from each cell population and repertoire analysis was performed. carried out. The pie chart shown in Figure 7(A) is the result of TCR repertoire analysis for each cell population, and T cells expressing the relevant TCR were observed in 3 cells in the PD-1 ⁺ 4-1BB ⁺ fraction. , which indicates that it was observed only in one cell (single) in the PD-1 ⁺ 4-1BB ⁻ fraction.
Regarding the reactivity of TCR to mutated antigens, we created four types of TMG (TM1 to TM4), introduced the mRNA created from each TMG into CCP15 patient-derived LCL, and analyzed it by ELISPOT. It was confirmed that T cells transfected with TMG1 showed reactivity to TMG1 (Figure 7(B)). Further analysis revealed that it specifically recognized mutant ABCF1 in TMG1 (Figure 7(C). 6 in TM1 is "ABCF1"). The amino acid sequences in the figure are "TKQAEKQTKEALTQKQQKCRRKNQDEE" (SEQ ID NO: 11) for mutant ABCF1 and "TKQAEKQTKEALTRKQQKCRRKNQDEE" (SEQ ID NO: 12) for wild type ABCF1.

Figure 8 shows a summary of the characteristics of all tumor-reactive TCRs derived from 4 cases (CCP1, 3, 5, 15).
All tumor-reactive TCRs were most frequently identified in the PD-1 ⁺ 4-1BB ⁺ DP population. Moreover, all variant antigen-specific TCRs, including the TCR at CCP5, were identified in cases with a high frequency of PD-1 ⁺ 4-1BB ⁺ population in TILs. Figure 8(B) shows the analysis results using CCP1. It was thought that tumor-reactive T cells, including CCP3,15, were more frequently enriched in the PD-1 ⁺ 4-1BB ⁺ fraction than in the PD-1 ⁺ 4-1BB ^- fraction. . In addition, none of the TCRs for which recognition mutant antigens could be identified were found in the PD-1 ^- fraction.
Figure 9 shows the log rank of the prognosis of the 16 analyzed colorectal cancer cases based on the "PD-1 ⁺ 4-1BB ⁺ cell population frequency in CD8 ⁺ TIL of 3% or more" shown in Figure 8 (A). The results of the test are shown. Although no significant difference was observed, no deaths were observed in the 4 relevant cases (CCP1, 3, 5, 15: corresponding to graph "A" in the figure) until 2000 days, while the remaining 12 cases Among the cases (corresponding to graph "B" in the figure), there were 7 survivors after 2000 days. In this way, it was determined that cases with "PD-1 ⁺ 4-1BB ⁺ cell population frequency in CD8 ⁺ TIL of 3% or more" tended to have a favorable prognosis.
FIG. 10 shows the results of sc (single cell) RNA-seq analysis and TCR-seq analysis using CCP1 case TIL. The results showed that CD8 ⁺ T cells expressing two types of TCR showed similar RNA expression patterns (FIG. 10(B)). Next, we performed an RNA expression comparison between CD8 ⁺ T cells expressing these two types of tumor-reactive TCRs and other CD8 ⁺ T cells, and found that the expression of CD9 RNA was higher than that of these two types of tumor-reactive TCRs. was shown to be significantly higher in T cells expressing (Figure 10(C)).

Figure 11 shows the results of an analysis using 9 CCP cases, adding CD9 as an index to PD-1/4-1BB, and narrowing down TILs. In the CCP9 case, TCRs were obtained from individual CD8 ⁺ T cells belonging to the PD-1 ⁺ TIL cell fraction, and two types of TCRs among them were confirmed to be reactive with CTOS derived from the CCP9 case (Figure 11 (A)). Next, using cryopreserved TIL from nine CCP cases, we obtained individual TCR genes from the PD-1 ⁺ 4-1BB ⁺ CD9 ⁺ CD8 ⁺ TIL cell fraction and performed repertoire analysis. As a result, it became clear that two types of tumor-reactive TCRs were identified more frequently in the PD-1 ⁺ 4-1BB ⁺ CD9 ⁺ fraction (Figure 11(B)). The obtained results indicate that the CD9 molecule is useful for narrowing down tumor-reactive T cells, and by adding CD9 to PD-1/4-1BB as an indicator, it is possible to further narrow down tumor-reactive T cells. It was thought that it would be.

The CD9 molecule is a type of molecule known as a "tetraspanin." Tetraspanins are a family of transmembrane proteins that penetrate the cell membrane four times, and are also called Transmembrane 4 superfamily (TS4SF). In addition to CD9, other known tetraspanins include TSPAN2, CD151, CD53, CD37, CD82, CD81, and CD63.
To date, there have been no reports that molecules belonging to tetraspanins serve as indicators for the selection of tumor-reactive T cells. Therefore, using a mouse model, we investigated whether the expression of CD9 molecules could be an indicator of tumor reactivity. As a mouse model, we used a tumor-bearing mouse treatment model in which CD8 ⁺ T cells specific for mutant Snd1 (staphylococcal nuclease domain-containing protein 1) are induced (Non-Patent Document 4). In this model, the mouse fibroblast sarcoma-derived cell line CMS7 was used as a tumor-bearing cell line, and by treatment with anti-checkpoint antibodies (anti-PD-1 antibody, anti-CTLA-4 antibody, and anti-GITR antibody), 21 days after tumor-bearing. CD8 ⁺ T cells specific for the mutant Snd1 encoded by the CMS7 tumor genome were confirmed in the spleen cells of patients.

The test method is shown in Figure 12(A). CMS7 was subcutaneously administered to mice on day 0, and anti-checkpoint antibodies were intravenously administered on days 7, 9, and 11. Spleens were collected on day 21, and mutant Snd1-derived peptide (SYAPCRGEF (SEQ ID NO: 13)) or wild-type Snd1-derived peptide (SYAPRRGEF (SEQ ID NO: 14)) was added to 1 × 10 ⁶ spleen cells, and IFN-γICS By using the intracellular staining method, it is possible to measure CD8 ⁺ T cells that react to peptides and produce IFN-γ intracellularly. This reaction is specific to the mutant Snd1-derived peptide and is not observed with the wild-type Snd1-derived peptide or the addition of DMSO as a solvent control. Furthermore, in tumor-bearing untreated mice that did not use anti-checkpoint antibodies, the response to the mutant Snd1-derived peptide was extremely low (FIG. 12(B)). The percentage of mutant Snd1 (mSnd1) showing IFN-γ production in the antibody treatment group was 1.6%, which was an extremely large number. FIG. 12(C) shows that the use of the mutant Snd1 peptide (31-mer) and adjuvant (polyIC:LC) was effective in treating CMS7 tumor-bearing mice.

FIG. 13 shows the results of examining whether TILs express CD9 molecules.
First, we investigated whether mutant Snd1-specific CD8 ⁺ T cells observed in a CMS7 tumor-bearing anti-checkpoint inhibitor (ICI)-treated mouse model expressed CD9 molecules. On day 0, the CMS7 tumor line (1×10 ⁶ cells) was subcutaneously inoculated into the back of BALB/c mice to carry the tumor, and on days 7, 9, and 11, treatment with ICI was performed. The group not treated with ICI served as the control. The spleen was removed from the mouse on day 21 (FIG. 13(A)). A mutant Snd1-derived peptide or a control peptide was added to 1×10 ⁶ spleen cells, and mutant Snd1-specific CD8 ⁺ T cells were measured using the IFN-γICS method (intracellular staining method). As a result, as shown in Figure 13(B), in the ICI treatment group, production of IFN-γ specific to the mutant Snd1-derived peptide was observed in CD8 ⁺ T cells, and the majority of them were a cell population expressing CD9 molecules. (the two data on the right in the upper row). On the other hand, no IFN-γ-producing CD8 ⁺ T cells were observed in the control peptide addition group (two data on the right in the middle row), and the frequency of mutant Snd1-derived peptide-specific CD8 ⁺ T cells was extremely low in the ICI-untreated group. It was low (two data on the right side of the bottom row).

Next, in order to examine the possibility that the mutant Snd1-derived peptide increases the expression of CD9 molecules, spleen cells were prepared from three mice in the ICI treatment group, and CD8 ⁻ was isolated using a mouse CD8 isolation kit (manufactured by Miltenyi Biotec). Spleen cells and CD8 ⁺ T cells were obtained. A portion of the obtained CD8 cells was stained using PE-conjugated mouse CD9 antibody, and then CD8 ⁺ CD9 ⁺ T cell population and CD8 ⁺ CD9 ⁻ T cell population were determined using anti-PE beads (manufactured by Miltenyi Biotec). obtained. These three types of 1×10 ⁵ CD8 cell populations (whole CD8, CD8 ⁺ CD9 ⁻ , CD8 ⁺ CD9 ⁺ ) were cocultured with 1×10 ⁶ CD8 ⁻ spleen cells supplemented with the mutant Snd1 peptide, and the mutant The frequency of type Snd1 peptide-specific CD8 ⁺ cells was analyzed using ELISPOT method. As a result, the expression of CD9 molecules in mutant Snd1-specific CD8 ⁺ T cells was not increased by peptide stimulation, but the majority of mutant Snd1-specific CD8 ⁺ T cells in the spleen of tumor-bearing mice were CD9 ⁺ . This became clear (Figure 13(C)).

CD9 is a type of molecule that belongs to tetraspanins, and many other molecules also belong to tetraspanins. Among these, it is known that female mice in which CD9 or CD81 are knocked out have problems with cell fusion during fertilization and become infertile, and functional similarities between tetraspanin molecules have been reported. Therefore, regarding the CD81 molecule, we analyzed the expression of the CD81 molecule in CD8 ⁺ T cells specific for the mutant antigen Snd1.
The results are shown in FIG. It was revealed that the majority of mutant antigen Snd1-specific CD8 ⁺ T cells expressed CD81 molecules as well as CD9 (FIG. 14(C)). Thus, it was considered that tumor-reactive CD8 ⁺ T cells were present in the CD8 ⁺ CD9 ⁺ CD81 ⁺ T cell population in the spleen cells of tumor-bearing mice. To test this idea, we investigated CD8 ⁺ CD9 ⁺ CD81 ⁺ T cells in the spleens of ICI-untreated mice after CMS7 tumor bearing, where the frequency of mutant Snd1-specific CD8 ⁺ T cells is extremely low (1.9% among CD8 ⁺ T cells). TCR repertoire analysis was performed. The results are shown on the right side of FIG. 14(B). Among the TCRs of the 36 cells analyzed, two types of TCRs had the same TCR gene sequence: 3 and 2 types, respectively.

From the results of TCR repertoire analysis of the CD8 ⁺ CD9 ⁺ CD81 ⁺ T cell population in the spleen of ICI-untreated mice after CMS7 tumor bearing, two types of TCR with multiple TCR expression and a single TCR were obtained. The two types of TCRs were expressed and introduced into BALB/c mouse T cells using retroviral vectors. Mouse T cells (1×10 ⁵ cells) introduced with each of these TCRs and CMS7 tumor cells (5×10 ⁴ cells) were co-cultured for 24 hours, and mouse IFN-γ and TNFα in the supernatant were measured. In the experiment, a mutant Snd1-specific TCR (mSnd1 22-1 TCR) that specifically reacts with CMS7 was used as a positive control.
As a result, as shown in FIG. 15, it was revealed that 11-3 TCR, in which the same TCR was observed in three cells, specifically recognized CMS7 and released these cytokines. This reaction was specific to CMS7, and like CMS7, no cytokine release was observed in CMS5, a mouse fibroblast sarcoma-derived cell line.

Next, we performed TCR repertoire analysis of the CD8 ⁺ CD9 ⁺ CD81 ⁺ T cell population in the spleen of CT26 tumor-bearing mice. In order to examine whether the same results as CMS7 could be obtained with mouse tumor cell lines other than CMS7, we used the CT26 cell line, a mouse colon cancer-derived tumor cell line. As shown in FIG. 16(A), 1×10 ⁶ CT26 cells were subcutaneously inoculated into the back of BALB/c mice on day 0, and CD8 ⁺ CD9 ⁺ CD81 ⁺ T cell population in spleen cells was isolated on day 21. TCR repertoire analysis was performed.
As shown in FIG. 16(B), as a result of analysis of the obtained 27 cell-derived TCRs, one type of TCR was expressed in two cells, and all the others were present in single form.

Next, one type of TCR in which the same TCR was observed in two cells, and two types of TCR in which the same TCR was present as a single (single) were selected, and their reactivity to CT26 tumor cells was verified.
First, expression of each TCR was introduced into BALB/c mouse T cells using a retroviral vector. These TCR-transfected mouse T cells (1 x 10 ⁵ cells) and CT26 tumor cells (5 x 10 ⁴ cells) were co-cultured for 24 hours, and mouse IFN-γ in the supernatant was measured. It became clear that one type of TCR (23-1 TCR) that was present specifically recognized CT26 (Figure 17(B)). It has been reported that CD8-positive T cell immune responses are elicited in CT26 tumor-bearing mice against the AH-1 peptide derived from mouse intrinsic leukemia virus envelope gp70, and the antigen recognized by the CT26 tumor-reactive TCR identified this time is AH-1. We verified the possibility that it was a peptide. Using P1.HTR cells whose mouse MHC (major histocompatibility antigen) is the same H-2d, TCR-transduced mouse T cells (1×10 ⁵ cells), AH-1 peptide addition, mutant Snd1 peptide addition, and related P1.HTR cells (5 × 10 ⁴ cells) with or without mutant ERK2-derived peptide were co-cultured for 24 hours, and mouse IFN-γ in the supernatant was measured (Figure 17(C) ).
As a result, it became clear that the identified antigenic peptide recognized by TCR was actually AH-1. The frequency of AH-1-specific CD8 ⁺ T cells among CD8-positive T cells in the spleen of CT26 tumor-bearing mice is assumed to be less than 1%, and these results indicate that the tumor-reactive CD8-positive T cells of CD9 and CD81 molecules, which are tetraspanin molecules, are expected to be less than 1%. This was considered to be extremely useful for narrowing down cells.

The functions of tetraspanins, including the CD9 molecule, are still not fully understood. This study has revealed that the expression of CD9 molecules plays an important role in the interaction between T cells and tumor cells. The possibility of increasing the effectiveness of therapy was considered. Therefore, in order to verify that CD9 expression enhances the antitumor effect in antigen-specific T cell infusion therapy, we obtained the following data.
A TCR transgenic mouse (DUC18, non-patent document 5) that specifically recognizes the complex of H-2Kd and the mutant antigen ERK2-derived peptide (mERK2-9m) encoded by the mouse fibrosarcoma-derived cell line CMS5 is used. did.
As shown in FIG. 18(A), CMS5 (1×10 ⁶ cells) was subcutaneously inoculated on the back of BALB/c mice on day 0, and cell transfusion (ACT) was performed on day 3. The following three groups (n=4/group) were established as treatment groups.
(1) No treatment (control), (2) Infusion of T cells (2 x 10 ⁶ cells) stimulated and proliferated for 5 days using anti-CD3 antibody and anti-CD28 antibody to DUC18 mouse spleen cells, (3) DUC18 Anti-CD3 and anti-CD28 antibodies were used in mouse spleen cells, and the next day, the mouse CD9 gene was introduced using a retrovirus, followed by infusion of T cells (2 x 10 ⁶ cells) that had been stimulated and proliferated for 4 days. .
As a result, as shown in FIG. 18(B), it was shown that the therapeutic effect was enhanced by infusion of T cells strongly expressing CD9 molecules.
As described above, according to this embodiment, T cells for tumor treatment of cancer patients could be provided. These T cells can be effectively used to treat cancer patients.

Claims (4)

(1) From tumor-infiltrating T cells (TILs) or whole blood collected from cancer patients, CD-8 positive T cells were detected using PD-1 and/or 4-1BB among cell surface antigens as markers. a candidate isolation step of isolating cells; (2) transducing the T cell receptor (TCR) gene of the T cells isolated in the candidate isolation step into target cells to generate TCR with tumor reactivity; Tumor reactivity evaluation step to identify and evaluate, (3) therapeutic T to proliferate T cells with TCR that reacts with the tumor in vitro to obtain T cells for tumor treatment to be administered to the cancer patient. A method for producing tumor-reactive T cells, comprising a cell obtaining step. The solid tumor is brain tumor, tongue cancer, esophageal cancer, stomach cancer, small intestine cancer, colon cancer, liver cancer, kidney cancer, bladder cancer, lung cancer, thyroid cancer, breast cancer, uterine cancer, or ovarian cancer. The method for producing tumor-reactive T cells according to claim 1, which is at least one selected from the group consisting of cancer, prostate cancer, rhabdomyosarcoma, and leiomyoma. 3. The method for producing tumor-reactive T cells according to claim 1, further comprising using tetraspanin as a marker in the candidate isolation step. 4. The method for producing tumor-reactive T cells according to claim 3, wherein the tetraspanin is at least one selected from the group consisting of CD9, TSPAN2, CD151, CD53, CD37, CD82, CD81, and CD63.