The present invention provides a method for predicting effect of immunotherapy on a cancer patient and/or prognosis of the cancer patient after the immunotherapy, and a gene set and a kit for use in the method. The present application claims the benefit of Japanese patent application number 2009-230279 and the subject matter of which is hereby incorporated herein by reference.
Cancer immunotherapy is still not an optimal therapeutic option for all patients although it has been effective for some patients. One of the reasons is that immunotherapy relys on immunity which considerably varies among different individuals to suppress proliferation of cancer cells. At present, it is not possible to predict effect of cancer immunotherapy and thus effectiveness of the therapy cannot be determined unless the therapy has done. It has been known that effect of chemotherapy on a breast cancer patient is predicted by determination of gene expression level. However, this prediction method is complicated because it involves not only gene expression level but also other factors. In addition, this method only directs to chemotherapy on breast cancer (patent literature 1). So far, methods for predicting effect of cancer immunotherapy or prognosis of a patient after immunotherapy have not been known.
An object of the present invention is to provide a method for predicting effect of immunotherapy on a cancer patient with accuracy.
Based on the results of peptide vaccine therapy on prostate cancer patients over the years, the present inventors tried to predict effect of cancer immunotherapy. First, gene expression profiles of prostate cancer patients before peptide vaccine therapy were analyzed with DNA micorarray. Then, the patients were classified into good prognosis group and poor prognosis group according to their survival time after the therapy, and correlation between expression level of each gene before the therapy and survival time after the therapy was analyzed. As a result, some genes were found to have correlation between the two. Finally, it was confirmed that survival time of a patient after immunotherapy could be predicted from expression level of those genes and the present invention has been accomplished.
Thus, the present invention provides a method for predicting effect of immunotherapy on a cancer patient and/or prognosis of the cancer patient after the immunotherapy, which comprises the steps of:
(1) determining expression level of each gene in a gene set which consists of at least 1 gene selected from the group of genes shown in any of Tables 1-5 and
(2) performing statistical processing of the expression level to calculate predicted survival time.
The present invention further provides a gene set for predicting effect of immunotherapy on a cancer patient and/or prognosis of the cancer patient after the immunotherapy, which consists of at least 1 gene selected from the group of genes shown in any of Tables 1-5.
The present invention further provides a kit for predicting effect of immunotherapy on a cancer patient and/or prognosis of the cancer patient after the immunotherapy, which comprises a probe or primer for each gene in the aforementioned gene set.
According to the present invention, it has become possible to predict effect of immunotherapy on a cancer patient and/or prognosis of the cancer patient after the immunotherapy by determining gene expression profile of the cancer patient before initiation of the immunotherapy. The present invention can offer useful information in selecting therapeutic options for cancer patients.
The prediction method of the present invention uses a gene or genes of which expression levels closely correlate survival time of a cancer patient after immunotherapy. The gene set of the present invention consists of at least 1 gene selected from the group of genes shown in any of Tables 1-5. The genes may be selected from any of Tables 1-5 in descending order from the top (the first gene), in ascending order from the bottom (the 300th gene), or randomly. In view of prediction accuracy, the genes are preferred to be selected in descending order from the top of any of the tables.
Number of genes in a gene set is not specifically limited and may be 1 or more. For example, for a gene set, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240, 250, 260, 270, 280, 290, or 300 genes may be selected.
The cancer patient to be treated by the present invention may be, but not limited to, a cancer patient of prostate cancer, pancreatic cancer, breast cancer, or liver cancer. Particularly, the present invention is suitable for a prostate cancer patient.
In the present invention, immunotherapy refers to a therapeutic method for treating cancer by stimulating immune response in a cancer patient against a tumor antigen protein. The immunotherapy of the present invention may include peptide vaccine therapy which uses a tumor antigen peptide, adoptive immunotherapy which uses lymphocytes such as cytotoxic T cells or natural killer cells, DNA vaccine therapy wherein a virus vector expressing a tumor antigen protein or peptide is introduced into a living body, and dendritic cell vaccine therapy wherein dendritic cells presenting a tumor antigen peptide are administered. Particularly, the present invention is suitable for peptide vaccine therapy.
In the present invention, expression level of a gene may be determined by any conventional method. Methods for determining expression level of a gene may include DNA microarray analysis, DNA chip analysis, PCR analysis, and Northern blot analysis, and particularly, DNA microarray analysis is suitable for the present invention.
DNA microarray analysis uses a microarray containing a probe for a gene to be determined. The microarray may be HumanWG-6 v3.0 Expression BeadChip (llumina). Alternatively, it is possible to prepare any desired microarray by synthesizing a probe for a gene to be determined and immobilizing the probe on a suitable substrate such as a slide glass. Methods for preparing a microarray are well known in the art. Analysis of microarray data is also well known in the art and may be performed according to MICROARRAYS FOR AN INTEGRATIVE GENOMICS, translated by Yujin Hoshida, Springer-Verlag Tokyo, for example.
A patient-derived sample used for determination of expression level of a gene may be, but not limited to, peripheral blood collected from a patient.
Expression level of a gene may be determined by DNA microarray analysis as follows. Briefly, total RNA is extracted and purified from peripheral blood of a patient. Then, biotinated cRNA is synthesized by using, for example, Illumina TotalPrep RNA Amplification Kit (Ambion). The biotinated cRNA is hybridized to a microarray and reacted with Cy3-labeled streptavidin. The resulting microarray is scanned by a dedicated scanner and Cy3 fluorescence of each spot is quantified by a dedicated software such as BeadStudio. Accordingly, expression level of the gene may be determined.
Predicted survival time of a patient may be calculated, for example, by applying the expression level of each gene of the gene set of the present invention in the patient into the regression equation of PLS regression analysis as shown in the following examples. Alternatively, it may be calculated by determining expression level of each gene of any gene set of the present invention in the patient group of the examples or other patient groups and performing statistical processing by SVM (Support Vector Machine), Regularized Least Squares, or Principal Component Analysis.
The gene set of the present invention is used for predicting effect of immunotherapy on a cancer patient and/or prognosis of the cancer patient after the immunotherapy and may be used for preparing a probe for DNA microarray or a primer for PCR for use in the prediction method of the present invention.
The kit of the present invention comprises a probe or primer which is capable of determining expression level of each gene of the gene set of the present invention. The probe or primer for each gene may be synthesized by a conventional method based on the sequence of the gene. The kit may comprise other reagents required for determination of expression level. The kit of the present invention may be used for DNA microarray analysis, DNA chip analysis, PCR analysis, and Northern blot analysis. A kit for DNA microarray analysis may comprise a microarray containing the probe as described above immobilized on a suitable substrate.
Patient-derived samples were peripheral blood which had been collected from prostate cancer patients who gave informed consent under the protocol approved by Kurume University ethical committee when the patients were diagnosed as recurrent prostate cancer in the past clinical trials. Gene expression profiles of 40 prostate cancer patients before peptide vaccine therapy were analyzed by using DNA microarray (HumanWG-6 v3.0 Expression BeadChip (Ilumina)). The prostate cancer patients included 20 good prognosis patients (whose survival time after peptide vaccine therapy was 700 days or more) and 20 poor prognosis patients (whose survival time after peptide vaccine therapy was less than 700 days) (
(I) RNA Extraction and Purification from Peripheral Blood of Patients
1. To peripheral blood sample of a patient, TRIzol LS (Invitrogen) was added at a ratio 1:3 and mixed such that the mixture became turbid.
2. To 750 μl TRIzol LS solution, 200 μl chloroform was added and mixed such that the mixture became turbid, and the resulting solution was centrifuged.
3. The supernatant obtained was removed to a fresh tube and added with ethanol 0.55 times as much as the supernatant.
4. The sample obtained in item 3 was mounted on a column of SV Total RNA Isolation System (Promega) and passed through the filter.
5. The filter was washed with 500 μl washing buffer.
6. Total RNA was eluted with 80 μl nuclease-free water.
7. Concentration of RNA was determined by using a spectrophotometer, and quality of RNA was checked by electrophoresis by using Experion system (Biorad).
(II) cRNA Synthesis for Microarray Using Illumina TotalPrep RNA Amplification Kit (Ambion)
(1) Synthesis of Single-Stranded cDNA by Reverse Transcription
1. To 500 μg total RNA each, nuclease-free water was added up to 11 μl.
2. The solution obtained item 1 was added with 9 μl Reverse Transcription Master Mix and incubated for 2 hours at 42° C.
1. To the tube of item (1)-2, 80 μl Second Strand Master Mix was added.
2. The tube was incubated for 2 hours at 16° C.
(3) Purification of cDNA
1. To the tube, 250 μl cDNA Binding Buffer was added.
2. The solution obtained in item 1 was mounted on cDNA Filter Cartridge and passed through the filter by centrifugation.
3. The filter was washed with 500 μl washing buffer.
4. cDNA was eluted with 19 μl nuclease-free water which was pre-heated to 50-55° C.
(4) Synthesis of cRNA by Reverse Transcription Reaction In Vitro
1. To the cDNA sample obtained in item (3)-4, 7.5 μl IVT Master Mix was added.
2. The tube obtained in item 1 was incubated for 14 hours at 37° C.
3. To the tube obtained in item 2, 75 μl nuclease-free water was added.
(5) Purification of cRNA
1. To the tube, 350 μl cRNA Binding Buffer was added.
2. To the tube, 250 μl of 100% ethanol was added and mixed such that the mixture became turbid.
3. The sample obtained in item 2 was mounted on tRNA Filter Cartridge and passed through the filter by centrifugation.
4. The filter was washed with 650 μl washing buffer.
5. cRNA was eluted with 100 μl nuclease-free water which was pre-heated to 50-55° C.
6. Concentration of cRNA of the resulting solution was determined and the solution was used as a hybridization sample.
(1) Preparation of cRNA for hybridization
1. To 500 μg total RNA, nuclease-free water was added up to 10 μl.
2. The solution obtained in item 1 was added with 20 μl GEX-HYB and incubated for 5 minutes at 65° C.
1. cRNA sample thus prepared was applied to HumanWG-6 v3.0 Expression BeadChip placed in a dedicated chamber.
2. The dedicated chamber was covered and incubated for 18 hours at 55° C.
1. The covers of the microarrays were removed in Wash E1BC solution.
2. The arrays were immediately set in a slide rack and washed with 1×High-Temp Wash buffer which was pre-heated to 55° C.
3. The arrays were washed in Wash E1BC solution for 5 minutes.
4. The arrays were washed in ethanol for 5 minutes.
5. The arrays were washed in Wash E1BC solution for 5 minutes.
6. To each of staining trays 4 ml Block E1 buffer was added, and then the arrays were set in the staining trays one by one to perform blocking for 10 minutes at room temperature.
7. To each of staining trays 2 μl streptoavidin-Cy3 per 2 ml Block E1 buffer was added, and then the arrays were set in the staining trays one by one to perform staining for 10 minutes at room temperature.
8. The arrays were washed in Wash E1BC solution for 5 minutes and dried by centrifugation.
1. The arrays were set in a dedicated scanner (Illumina) and scanned in a standard mode.
2. After scanning, each spot on the microarrays was quantified by using a dedicated software BeadStudio.
The microarray data obtained was normalized with VST (Variance Stabilizing Transformation) and RSN (robust spline normalization). Expression level of a gene was determined to be significant when the gene showed Presence Probability &#60 0.05 compared to a negative control (expression level measured with a probe for a gene which was not present in the microarray) (
The prostate cancer patients of Example 1 were classified into 2 groups, good prognosis group (survival time after peptide vaccine therapy was 700 days or more) and poor prognosis group (survival time after peptide vaccine therapy was less than 700 days) and genes were selected based on correlation between expression level of each gene and survival time, and expression variation between the two groups. Expression variation was represented as increase or decrease of mean expression level of poor prognosis group compared to that of good prognosis group as a control.
For analysis, the following five statistical methods were used (A: method for analysis of correlation between expression level of a gene and survival time, B: method for analysis of expression variation between good prognosis group and poor prognosis group).
Those references are herein incorporated by reference.
PLS (Partial Least Square) regression model was built for prediction of survival time of each patient after peptide vaccine therapy. Specifically, from each of the five gene sets consisting of top 300 genes selected by one of the above five methods, respectively (Tables 1-5,
Analysis of the regression model was performed according to Leave One Out Cross Validation (LOOCV) method. Namely, for 40 prostate cancer patients in total, regression model was built using results of 39 patients, and the regression model thus built was used to predict survival time of the remaining 1 patient.
For example, calculation of survival time with a gene set consisting of top 50 genes is as follows. Predicated survival time was calculated according to the following regression equation.
Predicated survival time=a1g1+a2g2+a3g3+ . . . +a50g50+E (formula 1),
wherein each of a1, a2, a3, . . . and a50 is a coefficient for expression level of a gene used to build the regression model; each of g1, g2, g3, g4, . . . g50 is expression level of a gene in a patient to be analyzed; and E is a constant term of the regression equation).
Table 6 shows coefficients used for the prediction of survival time of patients with the gene set consisting of top 50 genes selected by Pearson product-moment correlation coefficient, latent variable 3.
Those references are herein incorporated by reference.
The answer was determined to be correct when the predicted survival time was 700 days or more for a good prognosis patient or less than 700 days for a poor prognosis patient, and accuracy rate (prediction accuracy) of the regression model was calculated. When evaluated, all the regression models showed the best prediction accuracy more than 80% (
In addition, Kaplan-Meier curve was prepared using actual survival time of good and poor prognosis groups classified with the gene set consisting of: top 50 genes selected by Pearson product-moment correlation coefficient, latent variable 3 and then log rank test was performed. As a result, it was revealed that the present method could classify the patients into good and poor prognosis groups with probability p=1.6e-10 (
Then, from the gene set of Table 1, 30 gene sets each consisting of genes in multiples of 10 (i.e., 10, 20, . . . or 300 genes) in ascending order from the bottom were extracted and PLS regression model was built with latent variable 3 (
Gene Expression data of 9 prostate cancer patients were further included in addition to those of 40 prostate cancer patients of Example 1, and the patients were classified into good prognosis group (survival time was 300 days or more) and poor prognosis group (survival time was less than 300 days). Then, PLS regression model was built with a gene set consisting of top 50 genes selected by Pearson product-moment correlation coefficient, latent variable 3 and survival time of the 9 prostate cancer patients newly included were predicted (
For 40 prostate cancer patients of Example 1, a primary linear regression equation was prepared according to their survival time and expression level of one of top 300 genes selected by Pearson product-moment correlation coefficient (Table 1). Then, predicted survival time of a patient was calculated with the equation thus prepared and expression level in the patient, and the calculated predicted survival time (y axis) and actual survival time (x axis) were shown in a graph (
indicates data missing or illegible when filed
Drosophila); translocated to, 10”
indicates data missing or illegible when filed
Drosophila)”
Number | Date | Country | Kind |
---|---|---|---|
2009-230279 | Oct 2009 | JP | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/JP2010/067088 | 9/30/2010 | WO | 00 | 3/28/2012 |