JP2017118841A - Effect prediction marker for immunotherapy - Google Patents

Abstract

PROBLEM TO BE SOLVED: To develop a clinical effect prediction method for a cancer immunotherapy that is technically simple, is inexpensive, and has excellent sensitivity and accuracy.SOLUTION: The present invention provides a method and a kit for predicting a clinical effect of a cancer immunotherapy. The method and kit predict a clinical effect of a cancer immunotherapy to be higher as the expression level of the syndecan-4 gene or protein in a sample obtained from an object is lower.SELECTED DRAWING: None

Description

  The present invention relates to a method for predicting the clinical effect of immunotherapy for cancer and a kit therefor.

  In recent years, cancer immunotherapy has become a standard treatment option. In order to increase the effectiveness of immunotherapy, not only the biological activity is enhanced, but also an index (biomarker) that predicts effective / ineffective before inoculation is important.

  Cancer antigen peptides are often used for cancer immunotherapy. WT1 peptide which is one of cancer antigen peptides is a part of WT1 protein encoded by WT1 gene (Wilms' tumor gene 1). Currently, various cancer immunotherapy is performed using WT1 peptide. So far, there are several reports regarding predicting the clinical effect of cancer immunotherapy (WT1 vaccine) using WT1 peptide. Malignant glioblastoma patients with high WT1 expression levels have been reported to survive longer than those with low WT1 expression levels (Non-Patent Document 1). It has been reported that the frequency of higher WT1-specific CTL (cytotoxic T lymphocytes) is an indicator that the WT1 vaccine is successful (Non-Patent Documents 2 to 4). Reported that increased frequency of effector memory subsets (EMs), which are important for maintaining WT1-specific CTL, during repeated treatment of WT1 vaccination is also a predictor of WT1 vaccine efficacy (Non-Patent Document 4). Peripheral blood lymphocyte counts and delayed hypersensitivity reactions (DTH) to tumor antigens after immunization are also reported to be prognostic factors for clinical efficacy as reported in various immunotherapy (Non-Patent Literature). 5). However, in order to predict the clinical effect based on the above report, there are cases where there are problems in terms of technical difficulty, cost, and prediction sensitivity.

Chiba Y. et al. Brain tumor pathology, 27 (1), 29-34 (2010) Oka Y TA. Et al. Proc Natl Acad Sci U.S.A., 101 (38), 13885-13890 (2004) Hashii Y. et al. Leukemia, 26 (3), 530-532 (2012) Nishida S. et al. Journal of immunotherapy (Hagerstown, Md .: 1997), 37 (2), 105-114 (2014) Ogi C. et al. Immunotherapy, 6 (10), 1025-1036 (2014)

  The problem to be solved by the present invention was to develop a method for predicting the clinical effect of cancer immunotherapy that is technically simple, inexpensive, and excellent in sensitivity and accuracy. Specifically, the problem to be solved by the present invention was to find an appropriate biomarker for selecting patients for whom immunotherapy of cancer is expected to be effective.

  The present inventors conducted comprehensive gene expression analysis of peripheral blood mononuclear cells of brain tumor patients who received immunotherapy of cancer using WT1 peptide as an example of cancer antigen peptide. The present inventors have found that low expression correlates with clinical effects and have completed the present invention. Thus, the present invention provides the following.

(1) A method for predicting the clinical effect of immunotherapy of cancer, comprising measuring the expression level of syndecan-4 gene or protein in a sample obtained from a subject, and comprising syndecan-4 gene or A method of predicting that the lower the protein expression level, the higher the clinical effect.
(2) The method according to (1), wherein the expression level of the syndecan-4 gene is compared with the expression level of the internal standard gene.
(3) When the internal standard gene is beta-actin and the expression level of syndecan-4 gene is 1 or less when the expression level of beta-actin is 1, it is predicted that the clinical effect is high ( 2) The method described.
(4) The method according to any one of (1) to (3), wherein the sample is a peripheral blood sample.
(5) A kit for predicting the clinical effect of immunotherapy for cancer, comprising a reagent for measuring the expression level of syndecan-4 gene or protein, the lower the expression level of syndecan-4 gene or protein A kit that predicts high clinical efficacy.
(6) The kit according to (5), comprising a reagent for measuring the expression level of the internal standard gene.
(7) When the internal standard gene is beta-actin and the expression level of syndecan-4 gene is 1 or less when the expression level of beta-actin is 1, it is predicted that the clinical effect is high ( 6) The kit according to the above.
(8) The kit according to any one of (5) to (7), wherein the sample is a peripheral blood sample.

  According to the present invention, clinical effects of cancer immunotherapy can be predicted before treatment. The method of the present invention is technically very simple and inexpensive because the clinical effect can be predicted only by examining the expression of one gene or protein. The method of the present invention also has high prediction sensitivity and prediction accuracy.

FIG. 1 is a diagram for explaining selection of candidate genes by DNA microarray analysis. FIG. 1a is a diagram illustrating a method for finding a biomarker of the present invention. First, quantitative RT using 30 glioblastoma patients in the discovery set was found that genes with differential expression (DEGs) were screened by a cDNA microarray, and that the expression levels of the screened genes could be determined by different measurement methods. -Confirmed by PCR. The confirmed DEGs were then validated using 23 glioblastoma patients in the validation set (different from those in the discovery set). Finally, syndecan-4 was selected as a biomarker. FIG. 1b shows a Volcano plot. Each dot corresponds to one gene. X-axis and Y-axis are the logarithmic value of the change rate of signal intensity between long-term survivors and short-term survivors (log ₂ [short / long]), and between long-term survivors and short-term survivors, respectively. The logarithmic value (-log ₁₀ [P value]) of the statistically significant difference of the signal intensity difference of each gene is shown. Dashed lines indicate y = | x | -1, y = 1 and | x | = 0.5. FIG. 1c is a Volcano plot for 32 candidate DEGs extracted as described in the examples. Table 3 shows gene names and statistical evaluation. FIG. 2 is a diagram showing the effect of syndecan-4 (SDC-4) as a biomarker of the present invention. FIG. 2a shows the ROC curve when the OS (overall lifetime) threshold is chosen to maximize AUC (area under the curve). A cut-off value of 0.001 for the SDC-4 expression level can best distinguish long-term survivors (OS ≧ 256 days) and short-term survivors (OS <256 days). Long term survivors have an SDC-4 expression level of 0.001 or less. Biomarker performance is 70.4% sensitivity, 76.0% specificity, 76.0% positive predictive value, 70.4% negative predictive value, and 73.1% accuracy, The chi-square test showed a significant difference (P <0.001). This analysis used all patients in the discovery set and validation set. FIG. 2b shows Kaplan-Meier curves for OS (days) of SDC-4 low expression (0.001 or less) and SDC-4 high expression (greater than 0.001) patients. When the two groups were compared using a log rank test, a statistically significant difference in OS was observed between the two groups (p <0.001). The one-year survival rates were 64.0% and 18.5% for SDC-4 low expression patients and SDC-4 high expression patients, respectively. FIG. 3 is a diagram showing protein expression in peripheral blood mononuclear cells of syndecan-4 (SDC-4) as a biomarker of the present invention. MFI is the average fluorescence intensity. Black bars indicate MFI for long-term survivors and white bars indicate short-term survivors.

  The present invention, in the first aspect, is a method for predicting the clinical effect of immunotherapy for cancer, wherein the expression level of syndecan-4 gene or protein in a sample obtained from a subject is measured. In addition, the present invention relates to a method for predicting that the lower the expression level of a syndecan-4 (SDC-4) gene or protein, the higher the clinical effect. The cancer immunotherapy may use any cancer antigen, and examples include cancer immunotherapy using the WT1 peptide.

When administering a target antigen used for immunotherapy of cancer whose clinical effect is predicted according to the present invention, it is a peptide containing a partial amino acid of the gene product that becomes the antigen, or a modified peptide thereof, and administered to the subject as an immunogen Any peptide may be used as long as it has anticancer activity. Various peptides (cancer antigen peptides) that can be used for cancer immunotherapy are known, and one example is the WT1 peptide. Such WT1 peptides include modified WT1 ₂₃₅ peptide (SEQ ID NO: 1), WT1 ₂₃₅ peptide (SEQ ID NO: 2), WT1 ₁₂₆ peptide (SEQ ID NO: 3), WT1 ₁₈₇ peptide (SEQ ID NO: 4), Examples include, but are not limited to, WT1 ₃₃₂ helper peptide (SEQ ID NO: 5).

  Cancer immunotherapy encompasses treatment and prevention. Therefore, the subject may be a human who has cancer, a human who may have cancer, or a healthy human. Whether or not cancer immunotherapy is effective in a subject scheduled for cancer immunotherapy may be predicted by the present invention. Whether or not cancer immunotherapy is effective in subjects undergoing other cancer treatments may be predicted by the present invention. Alternatively, the present invention may predict whether it is effective to prevent cancer by immunotherapy in a subject.

  Whether a cancer antigen peptide exerts a therapeutic or preventive effect on cancer in a subject depends on whether the cancer antigen peptide corresponds to the HLA type of the subject. Since it is known which HLA type is compatible with many cancer antigen peptides, the cancer antigen peptide used in the present invention can be selected according to the HLA type of the subject. From this point of view, the subject to which cancer immunotherapy with a cancer antigen peptide is applied may be any person with any HLA type.

  The cancer subject to immunotherapy may be any type of cancer, for example, leukemia such as chronic myelogenous leukemia, hematopoietic tumor such as myelodysplastic syndrome, multiple myeloma, malignant lymphoma, or esophageal cancer. , Stomach cancer, colon cancer, pancreatic cancer, lung cancer, breast cancer, germ cell cancer, liver cancer, biliary tract cancer, head and neck cancer, skin cancer, sarcoma, kidney cancer, bladder cancer, prostate cancer, testicular cancer, uterine cancer, cervical cancer, Examples include, but are not limited to, ovarian cancer, thyroid cancer, carcinoid, pulmonary blastoma, hepatoblastoma, brain tumor or thymic cancer.

  In practicing the present invention, the expression level of the SDC-4 gene or protein in a sample obtained from the subject is examined. The sample may be collected from any tissue or site of the subject, and is preferably a peripheral blood sample, and the expression level of SDC-4 gene or protein in lymphocytes or mononuclear cells therein is examined. It is preferable.

  In order to examine the expression level of the SDC-4 gene or protein, known gene / protein expression level measurement methods such as Northern blotting, PCR, Western blotting, flow cytometry, ELISA, etc. can be used. Preferably, a quantitative PCR method, particularly preferably a quantitative real-time PCR (RT-PCR) method is used. The quantitative RT-PCR method is widely used by those skilled in the art, has high sensitivity and specificity, and can be quantified even with a small amount of gene expression level.

  In the present invention, it is predicted that the lower the expression level of syndecan-4 gene or protein in the sample obtained from the subject, the higher the clinical effect of cancer immunotherapy with a cancer antigen peptide. Preferably, it is determined whether the expression level of the syndecan-4 gene is low by comparing the expression level of the SDC-4 gene with the expression level of the internal standard gene.

  As the internal standard gene, a gene that does not vary in expression level depending on the tissue, experimental treatment or developmental stage is used. Examples of the internal standard gene include housekeeping genes such as beta-actin and glyceraldehyde 3-phosphate dehydrogenase (GAPDH), but are not limited to these genes.

  For example, when beta actin is used as an internal standard gene for measuring the expression level of the SDC-4 gene in a sample derived from the peripheral blood of the subject, if the beta actin expression level is 1, the expression level of SDC-4 is When it is 0.001 or less, it can be predicted that cancer immunotherapy with the WT1 peptide is effective in the subject. A person skilled in the art can appropriately determine the cut-off value in the present invention according to the internal standard gene and the type of sample.

  In another aspect, the present invention relates to a kit for predicting the clinical effect of cancer immunotherapy, comprising a reagent for measuring the expression level of SDC-4 gene or protein. Preferably, the expression level of the SDC-4 gene is measured using a quantitative RT-PCR method, particularly preferably using a quantitative RT-PCR method.

  As a reagent for measuring the expression level of the SDC-4 gene, for example, when quantitative PCR is used, there are primers for SDC-4 gene amplification and / or enzymes for performing quantitative RT-PCR. Although it is mentioned, it is not restricted to these. Preferably, the kit of the present invention may contain an internal standard gene amplification primer and / or an enzyme for performing quantitative PCR in addition to the SDC-4 gene amplification primer. Primers for SDC-4 amplification can be designed based on the known SDC-4 nucleotide sequence by those skilled in the art. Examples of primers for SDC-4 amplification using quantitative TR-PCR include, but are not limited to, GGCAGGAATCTGATGACTTTG (forward, SEQ ID NO: 52) and GCCGATCATGGAGTCTTCC (reverse, SEQ ID NO: 53). When beta actin is used as an internal standard gene, examples of beta actin amplification primers using quantitative TR-PCR include CCAACCGCGAGAAGATGA (forward, SEQ ID NO: 70) and CCAGAGGCGTACAGGGATAG (reverse, SEQ ID NO: 71). For example, but not limited to. As a reagent for measuring the expression level of SDC-4 gene or protein, it is used for the known gene / protein expression level measurement methods such as Northern blotting, other PCR methods, Western blotting, flow cytometry, ELISA, etc. It may be a reagent. Such a reagent can be appropriately selected and used by those skilled in the art.

  EXAMPLES Hereinafter, the present invention will be described more specifically and in detail with reference to examples. However, it should not be understood that the scope of the present invention is limited to the examples.

1. Materials and Methods (1) WT1 peptide vaccine WT1 peptide vaccine is composed of WT1-CTL epitope peptide and Montanide ISA51 adjuvant. This peptide used in this study is a modified 9-mer WT1 peptide (235-243 CYTWNQMNL (SEQ ID NO: 1); mWT1 ₂₃₅ ), a native WT1 ₂₃₅ peptide (235-243 CMTWNQMNL (SEQ ID NO: 2); Y in the second amino acid position, which is the anchor position for WT1 ₂₃₅ ) HLA-A ^* 24 ^: 02, is replaced by M. The binding affinity of mWT1 ₂₃₅ to the HLA-A ^* 24: 02 molecule is higher than WT1 ₂₃₅ , and mWT1 ₂₃₅ has been demonstrated to induce a much stronger CTL response against tumor cells expressing WT1 than WT1 _235. ing. GMP grade WT1 peptides were purchased from Multiple Peptide Systems (San Diego, California) and Peptide Institute (Osaka, Japan).

(2) Patients Sixty patients were enrolled in the phase II study of WT1 peptide vaccine against malignant glioblastoma (GBM). Patients with relapsed or advanced GBM were subjects enrolled in a phase II trial if the disease was resistant to conventional chemotherapy or radiation therapy. Other selection criteria were similar to our previous work. Briefly, 1) between ages 16 and 80, 2) WT1 expression in glioma cells determined by immunohistochemical analysis, 3) HLA-A ^* 2402 positive, 4) estimated survival 3 months or more, 5) ECOG performance status grade 0-2, 6) no severe organ dysfunction, and 7) written informed consent of the patient. All enrolled patients were histologically proven to have GBM (grade 4) based on WHO criteria.

  PBMC samples were available prior to WT1 vaccination from 53 of the 60 patients enrolled in the phase II trial. Of the 53 patients, 15 patients are long-term survivors with an overall survival (OS) ≧ 460 days (median: 1133 days, range: 480-2678 days) and the other 15 patients have OS Randomly selected as <460 days (median: 216 days, range: 138-458 days), where OS is the day from recurrence or disease progression to death or censored time Defined. These 30 patients (15 long-term survivors and 15 short-term survivors) are defined as patients in the discovery set and the remaining 23 patients are defined as patients in the validation set. It was done.

  A phase II trial of WT1 peptide vaccination and cDNA microarray analysis of blood samples was approved by the Ethics Review Board of Osaka University Hospital.

(3) WT1 peptide vaccination schedule After obtaining informed consent, intradermal injection of HLA-A ^* 24: 02 restricted mWT1 ₂₃₅ peptide emulsified with 3.0 mg Montanide ISA51 adjuvant was started. The schedule was to inoculate once a week for 12 consecutive weeks. Responses were in full remission (CR), partial remission (PR), stable (SD) using magnetic resonance imaging (MRI) at 12 weeks after initial vaccination according to RECIST (Solid Tumor Response Criterion) criteria. ) And exacerbation (PD). If clinical response becomes apparent, continue WT1 vaccination for the next few months, every 2 weeks, then every 1 to 3 months, until obvious tumor progression or worsening of the patient's health is observed did.

(4) Blood sample Peripheral blood was obtained from a patient, and then PBMC were separated by density gradient centrifugation using Lymphocyte Separation Solution (Nacalai Tesque, Kyoto, Japan). Separated PBMC were stored in liquid nitrogen until use.

(5) RNA isolation from PBMC and cDNA microarray analysis Thaw frozen PBMC, isolate total RNA using TRIzol reagent (Life technologies, Carlsbad, California), purify using chloroform, then isopropanol and ethanol Precipitation was performed. The amount of purified total RNA was measured using a nanoDrop ND-1000 (Thermo Fisher Scientific, Waltham, Massachusetts).

  RNA from PBMCs of patients in a discovery set was sent to an external manufacturer (Toray Industries, Tokyo, Japan) and RNA-based cDNA microarray analysis was performed using Human oligo chip 25k ver 1.00 (Toray Industries, Tokyo, Japan). ) And carried out there.

Microarray analysis data was transformed using global normalization followed by quantile normalization (Shippy R. et al. Nature biotechnology, 24 (9), 1123-1131 2006)). First, genes having an intensity of less than [(average intensity of blank spots) + 10 × (standard deviation of intensity of blank spots)] were excluded. To identify genes that are differentially expressed between long-term and short-term survivors (Differentially Expressed Gene, DEG), log ₂ (ratio between two groups or fold-change) (x-axis) and between two groups We created a volcano plot composed of -log ₁₀ (p value) (y-axis) of the Welch's t-test (Allison DB. Et al. Nature reviews. Genetics, 7 (1), 55-65 (2006)) . The logarithm of the ratio between the two groups was calculated by the following formula: log ₂ [(average signal intensity in long-term survivors) / (average signal intensity in short-term survivors)]. DEG was selected using the volcano plot according to the following conditions: y> | x | ⁻¹ , | x |> 0.5, and y> 1.0. To reduce the probability of false positives for DEG, the correlation between the expression intensity of individual genes and either progression-free survival (PFS) or treatment efficacy assessment according to the RECIST criteria, respectively, was compared with that of Pearson and Spearman. The correlation coefficient was investigated. Genes that did not show significant p-values (here alpha levels were set to 0.1 for the Pearson correlation coefficient and 0.2 for the Spearman correlation coefficient, a non-strict threshold) ) Was excluded from further analysis. Finally, the analysis of the next step was carried out with genes satisfying the following six conditions: 1) Average intensity of individual genes ≧ [(average intensity of blank spot) + 10 × (standard deviation of intensity of blank spot) )] ;; 2) -log ₁₀ (p-value of difference by t-test between long-term OS group and short-term OS group) = | log 2 (fold-difference) | ⁻¹ ; 3) | log 2 (double difference) |>0.5; 4) -log ₁₀ (p-value of difference by t-test between long-term OS group and short-term OS group)>1.0; 5) Pearson's correlation coefficient between PFS and microarray signal intensity p-value <0.1; and p-value <0.2 for Spearman's correlation coefficient between therapeutic effect determination and microarray signal intensity.

(6) Quantitative RT-PCR
RNA from PBMC was reverse transcribed into single-stranded cDNA using SuperScript VILO cDNA Synthesis Kit and Master Mix (Life technologies, Carlsbad, California). Primer sequences for RT-PCR are listed in Table 1. Target sequences are available in the NCBI nucleotide database (http://www.ncbi.nlm.nih.gov/nuccore). Prior to loading the sample cDNA into the BioMark 48.48 dynamic array nanofluidic chip (Fluidigm Corporation, San Francisco, Calif.), 15 cycles of pre-amplification were performed with the diluted primer mixture. RT-PCR was then performed using BioMark HD (Fluidigm Corporation, San Francisco, Calif.) According to the manufacturer's protocol. Beta actin (ACTB) was used as an internal control, and ΔCT was calculated as [(CT value of each gene) − (CT value of ACTB)]. All data on gene expression levels were converted to log ₂ (1 + 2 ^−ΔCT ) and ^subjected to statistical analysis.

(7) Quantification of protein expression by flow cytometry
Peripheral blood mononuclear of 3 patients with the longest survival and 3 patients with the shortest survival after vaccination among the patients with ovarian cancer that participated in the WT1 peptide vaccine phase II clinical trial, as in the GBM patient above Analysis was performed with a sphere. As the antibody, FITC-labeled anti-syndecan 4 antibody (5G9) (Santa Cruz, Dallas, Texas, USA) was used. Protein expression was evaluated by mean fluorescent intensity (MFI), and the value obtained by subtracting the MFI of the unstained specimen was used for analysis.

(8) Statistical analysis Statistical analysis of RT-PCR was performed using an OS defined as the survival time from the day of the start of WT1 vaccination to the time of death or censoring. The normal distribution of variables was evaluated using the Jarque-Bera test, and homogeneity of variation was confirmed using the F test.

  To assess differences in patient characteristics between long-term and short-term survivors in the discovery set and validation set, in age, gender, Karnovsky performance status (KPS), and radiation therapy The total absorbed dose, the history of surgical treatment, and the history of chemotherapy were analyzed using the Cox proportional hazard model.

  To examine the correlation between OS and individual gene expression levels measured by quantitative RT-PCR, logarithm, square root, cubic root, or fourth root so that each gene meets the normality requirement Appropriate conversion including conversion was performed. All normally distributed variables were checked for homogeneity of variance and analyzed by Student's t test. Variables that were not normally distributed were analyzed using the Mann-Whitney U test.

  The cut-off value of the SDC-4 expression level for predicting the treatment effective group / ineffective group is maximized as Youden's index (defined as Youden's J static = sensitivity + specificity-1) And determined using an ROC curve. The accuracy of predicting OS by sensitivity, specificity, positive and negative predictive value, and expression level of SDC-4 was calculated using the formula. The SDC-4 expression level cut-off value determined by the ROC curve was divided into two groups, an effective group and an ineffective group, and evaluated using a Kaplan-Meier curve and a two-sided log-rank test. Statistical analysis was performed using statistical software such as MP Pro 10.0 (SAS Institute, Inc. Cary, NC) and R-commander.

KPS, カルノフスキーのパフォーマンスステータス; RT, 放射線療法 (全照射線量) ;
SD, 標準偏差; OS, 全生存期間; PFS, 無増悪生存期間

†OSの起算日は腫瘍再発もしくは増悪した日
‡OSの起算日はワクチン接種を開始した日 2. Results (1) Using a discovery set, exploratory cDNA microarray analysis was performed to identify genes with differential expression in PBMC prior to WT1 vaccination.
Relapsed or previously refractory malignant glioblastoma patients who received the WT1 peptide vaccine were selected as cohorts for this analysis. PBMCs before WT1 vaccination were collected from 53 people who entered the vaccine clinical trials. An outline of a strategy for searching for biomarkers useful for predicting therapeutic effects is shown in FIG. 1a. Of the 53 patients, 30 were randomly registered in the discovery set and 23 were registered in the validation set. In the discovery set, 15 patients with OS ≧ 460 days (median: 1133 days, range: 480-2678 days) are long-term survivors, OS <460 days (median: 216 days, range) 138-458 days) were selected as short-term survivors (Table 2). Analysis using the Cox hazard model showed no association with treatment effect in age, sex, history of surgical treatment for brain tumors, or total absorbed dose of radiation therapy for brain tumors, but low performance status And the history of chemotherapy tended to be at high risk of reducing the therapeutic effect. All short-term survivors have died due to worsening of the primary disease at the time of censoring at the time of analysis (Table 3).

KPS, Karnovsky performance status; RT, radiation therapy (total dose);
SD, standard deviation; OS, overall survival; PFS, progression-free survival

† OS date is the date of tumor recurrence or worsening ‡ OS date is the date of vaccination

† log₂(長期/短期)，‡ two-tailed Welch's t-test CDNA microarray analysis was performed using PBMC prior to WT1 vaccination to find candidate genes that differ in expression between long-term and short-term survivors. log ₂ (ratio between two groups, ie, fold-change) is the x-axis, and logarithmic value of statistical significance (-log ₁₀ p value) is the y-axis. One is shown) in FIG. 1b. Subsequently, a gene (DEG) having a difference in expression was extracted as follows. First, genes with an intensity of less than [(average intensity of blank spots) + 10 × (standard deviation of intensity of blank spots)] are excluded, then the remaining 3,037 genes are classified into three criteria, y> | x According to | ⁻¹ , | x |> 0.5, and y> 1.0, the number was narrowed to 74. In order to further narrow down the candidate genes, the correlation between the signal intensity of each gene and either PFS or clinical response was assessed by Pearson and Spearman rank correlation analysis, respectively, and the former analysis was greater than 0.1 Genes with p-values and p-values greater than 0.2 for the latter analysis were excluded. Finally, 32 genes were selected as candidate genes (Table 4 and FIG. 1c). Of these 32 candidate genes, 11 were highly expressed in long-term survivors (Table 4, top), and 21 were highly expressed in short-term survivors (Table 4, bottom).

† log ₂ (long / short), ‡ two-tailed Welch's t-test

†log₂(長期生存／短期生存) (2) Verification of candidate genes selected by cDNA microarray by quantitative RT-PCR Expression levels of 32 candidate genes selected by cDNA microarray analysis in PBMCs derived from 30 discovery set patients Confirmed by quantitative RT-PCR.
In order to perform rigorous statistical analysis, the expression levels of individual genes obtained by quantitative RT-PCR are appropriately converted to meet the normality requirements, and the normal distribution is analyzed by the Jarque-Bera test. The homogeneity of variation was confirmed by F test, and the same trend of significance was confirmed by one-sided Student's t test (Table 5). If the normal distribution could not be converted, a one-sided Mann-Whitney U test was used. As a result of these statistical analyses, it was confirmed again that 15 out of 32 candidate genes were different between long-term survivors and short-term survivors (indicated as Verified in Table 5), and 17 genes Excluded (indicated as Excluded in Table 5).

† log ₂ (long-term survival / short-term survival)

KPS, カルノフスキーのパフォーマンスステータス; RT, 放射線療法 (全照射線量) ; SD, 標準偏差; OS, 全生存期間; PFS, 無増悪生存期間

†log₂(長期/短期)
‡表５に示したのと同じ統計学的方法を用いた。 (3) Validation of 15 candidate genes using validation set patients In order to confirm the reproducibility of whether these candidate genes are different between long-term survivors and short-term survivors even in different patient populations, 23 other malignant glioblastoma patients who were administered WT1 peptide vaccine with the same protocol as the patients in the discovery set were validated as a validation set. Table 6 shows the patient characteristics in the validation set. There was no confounding between OS and age, sex, performance status, presence or absence of surgery or chemotherapy, and total absorbed dose of previous radiation therapy. A validation set patient was used to confirm whether the expression levels of these 15 genes correlated with OS (Table 7). These analyzes were performed using the same statistical methods used in the discovery set. As a result, the SDC-4 gene expression level had a significant and negative correlation with OS (Table 7).

KPS, Karnovsky performance status; RT, radiation therapy (total dose); SD, standard deviation; OS, overall survival; PFS, progression-free survival

† log ₂ (long / short)
‡ The same statistical method as shown in Table 5 was used.

(4) Prediction of GBM patients treated with WT1 peptide vaccine by expression level of SDC-4 Prognosis of GBM patients treated with WT1 peptide vaccine using the expression level of SDC-4 in PBMC before WT1 vaccination The ability to predict was investigated.
First, receiver operating characteristic (ROC) curves were drawn using a total of 53 patients, a discovery set and a validation set (FIG. 2a). When the cut-off value of OS was set to 256 days, the “effective group” and “invalid group” could be most accurately predicted, and the area under the ROC curve (AUC) at that time was 0.72, which was the maximum value. The optimal cut-off value for predicting the effective group (OS is 256 days or more) and the ineffective group (OS is less than 256 days) is calculated to 0.001 by Youden's index, and 0.001 or less for the effective group In the ineffective group, it was greater than 0.001. The performance of the treatment effect prediction is 70.4% sensitivity, 76.0% specificity, 76.0% positive predictive value, 70.4% negative predictive value, and 73.1% accuracy, The chi-square test showed a significant difference (P <0.001).

  To further demonstrate the correlation between survival rate and expression level of SDC-4, patients were divided into 2 groups, SDC-4-low expression (SDC-4 ≦ 0.001) group and SDC-4-high expression (SDC). -4> 0.001) divided into groups (FIG. 2b). Thereafter, the difference in OS between the two groups was evaluated by Kaplan-Meier curve followed by log-rank test. The results showed that patients in the SDC-4-low expression group survived significantly longer than the SDC-4-high expression group. The annual OS rate was 64.0% and 18.5% for SDC-4-low-expressing patients and SDC-4-high-expressing patients, respectively.

KPS, カルノフスキーのパフォーマンスステータス; RT, 放射線療法 (全照射線量)
†時間変数として免疫療法開始からのＯＳを用いるコックスハザードモデルを使用した。 Furthermore, confounding with patient characteristics was examined using the Cox hazard model. Age and performance status are known prognostic factors in malignant glioblastoma, and the low performance status and previous history of chemotherapy in the discovery set tended to be a risk of shortening the prognosis. Can be adjusted for confounding with prognostic factors. As can be seen from the results shown in Table 8, high expression of SDC-4 was an independent clinical effect predictor, and significantly decreased the clinical effect. (Hazard ratio 13.8, 95% confidence interval 1.35-84.2, p = 0.027)

KPS, Karnovsky performance status; RT, radiation therapy (total dose)
† Cox hazard model using OS from start of immunotherapy as time variable was used.

(5) Quantification of protein expression by flow cytometry
Among the patients with ovarian cancer that participated in the WT1 peptide vaccine phase II clinical trial, analysis was performed on peripheral blood mononuclear cells of 3 patients with the longest survival time after vaccination and 3 patients with the shortest survival time. The cells were stained with fluorescently labeled anti-CD3 antibody, anti-CD4 antibody and anti-CD8 antibody, and the expression of SDC-4 protein in CD3 ⁺ T cells, CD4 ⁺ T cells and CD8 ⁺ T cells was analyzed by flow cytometry. For detection of SDC-4 protein, FITC-labeled anti-syndecan 4 antibody (5G9) (Santa Cruz, Dallas, Texas, USA) was used. Protein expression was evaluated by mean fluorescent intensity (MFI), and the value obtained by subtracting the MFI of the unstained specimen was used for analysis. In multi-color analysis, the expression of Syndecan-4 protein was low in long-term survivors in CD3 ⁺ T cells, CD4 ⁺ T cells and CD8 ⁺ T cells, which was consistent with the SDC-4 mRNA data. The results are shown in FIG.

  From the above experiments and analysis data, it was confirmed that the lower the expression level of the SDC-4 gene or protein, the higher the clinical effect of cancer immunotherapy.

  The present invention can be used in the fields of diagnostic agents, test agents, research reagents and the like.

Claims (8)

  A method for predicting the clinical effect of immunotherapy of cancer, comprising measuring the expression level of syndecan-4 gene or protein in a sample obtained from a subject, and expressing the expression of syndecan-4 gene or protein A method of predicting that the lower the level, the higher the clinical effect.   The method according to claim 1, wherein the expression level of the syndecan-4 gene is compared with the expression level of the internal standard gene.   The internal standard gene is beta-actin, and when the expression level of syndecan-4 gene is 0.001 or less when the expression level of beta-actin is 1, the clinical effect is predicted to be high. the method of.   The method according to any one of claims 1 to 3, wherein the sample is a peripheral blood sample.   A kit for predicting the clinical effect of immunotherapy for cancer, comprising a reagent for measuring the expression level of syndecan-4 gene or protein, and the lower the expression level of syndecan-4 gene or protein, the clinical effect A kit that predicts a high price.   The kit according to claim 5, comprising a reagent for measuring the expression level of the internal standard gene.   The internal standard gene is beta-actin, and when the expression level of syndecan-4 gene is 0.001 or less when the expression level of beta-actin is 1, the clinical effect is predicted to be high. Kit.   The kit according to any one of claims 5 to 7, wherein the sample is a peripheral blood sample.

Images