Patent Grant
12060617
The present invention relates to the field of classification, prognostics and treatment of cancer, in particular of prostate cancer.
The Sequence Listing submitted herewith as a text file named “37441_0017U2_Sequence_Listing,” created on Sep. 21, 2017, and having a size of 94,208 bytes is hereby incorporated by reference pursuant to 37 C.F.R. § 1.52(e)(5).
Accurate prognosis and prediction of overall and cancer specific survival at the time of prostate cancer diagnosis are of utmost importance for the improvement of current status of personalized treatment choice between radical prostatectomy, radiation therapy, castration therapy and watchful waiting (Shariat et al, Cancer 2008, 113:3075-3099; Touijer et al, Cancer 2009, 115:3107-3111; Freedland, Cancer 2011, 117:1123-1153). Radical prostatectomy and curative radiation therapy for men with localized prostate cancer can reduce mortality and prolong lives for patients with aggressive cancer. On the other hand for patients with less aggressive cancer, that potentially need no radical or curative treatments at all, such therapies may cause unnecessary complications and side effects. For patients with less aggressive cancer watchful-waiting or active surveillance can be a suitable option. However, current prognostic and predictive methods based on common clinical parameters—including age at diagnosis, serum PSA level, Gleason score of biopsies and clinical stage—cannot accurately distinguish between less aggressive and aggressive cancers at localized stage. Nor can they identify what kind of cancer can still be effectively controlled by castration therapy when the disease becomes dangerous.
For patients whose cancers are already advanced at diagnosis or have relapsed after curative treatments chemical or surgical castration can palliate symptoms and slow down disease progression. Unfortunately, the effect and side effects of castration therapy show strong variation among patients. Some can live longer than five years with minimal side effects whereas others can die of castration resistant metastasis within three years or die of cardiovascular and other side effects of the castration treatment. Currently there are no methods that can predict what kind of patients would be benefit most from castration therapy.
A majority of prostate cancers progress so slowly that they can never reach the life-threatening stage, mainly due to old age and other competing diseases. However, a small proportion of prostate cancers progress very rapidly and kill patients in less than five years. At diagnosis, by conventional clinical parameters including age, tumor grade, Gleason score, clinical stage and comorbidity the prediction of cancer specific and overall survival can reach an accuracy up to 60-70%. Even patients with the same clinical prognostic parameters can show strong difference in survival as well as in response to treatment. Hence prostate cancer is a pathological (morphorlogical) diagnosis that may include several different biological subgroups or subtypes.
There is a need for a method that can distinguish these biological subgroups or subtypes of prostate cancer patients. There is also a need for a method that can classify these subtypes into aggressive or high risk tumors and less aggressive or low risk tumors, as well as a method that can predict survival of the patients with the respective subtype tumor. Furthermore, there is a need for a method that can be used for making a treatment decision for patients that have a tumor of the respective subtype, possibly also taking into account clinical parameters.
Patent document WO2008/013492 A1 discloses an approach for identifying embryonic stem cell related genes, denoted ES tumor predictor genes (ESTP genes), that may be important for cancer stem cell function. 641 ESTP genes were identified and found useful for classification of prostate cancer tumors.
Patent document WO09021338 A1 discloses a method for prognosis of a cancer, e.g. prostate cancer, in a subject by detecting a signature of splicing events. F3 is mentioned as one of many genes that may be used.
Patent document WO0171355 discloses the simultaneous analysis of PSA, IGF-I and IGFBP-3 in blood plasma to predict the risk for a man of getting prostate cancer.
US2003054419 A1 discloses a method for determining the risk of progression of a prostate cancer patient after therapy, wherein the levels of TGF-β1, IGFBP-2, or IGFBP-3 in plasma are measured.
Patent documents WO10006048 A and US2009298082 A disclose methods for predicting survivability of prostate cancer diagnosed patient and whether a subject with PSA recurrence will later develop systemic disease respectively. In both disclosures IGFBP3 is mentioned as one of many genes that may be used, together with other molecular markers.
Documents WO09105154 and WO06028867A disclose a method for determining a prognosis for an individual having cancer and a method of diagnosis for multiple myeloma. c-MAF is mentioned as one of many genes that may be used.
WO10101888A discloses a method for interfering with the activity of CTGF, wherein the activity of CTGF is associated with prostate cancer metastasis.
It is an object of the present invention to provide molecular markers useful for classification, for prediction of prognosis and for guiding treatment decisions of prostate cancer in a subject.
It is another object of the present invention to provide new methods for classifying prostate cancer in a subject, as well as for using the classification for predicting prognosis of the subject and for making a treatment decision for the subject.
It is a further object of the present invention to provide a method of treating a subject having prostate cancer, based on the subject's tumor subtype.
It is still another object of the present invention to provide tools for classifying prostate cancers or tumors in a subject.
Identification of Genes and Gene Signatures that are Significantly Correlated to Survival in Prostate Cancer Subjects
To support the biological subtype concept, previous studies using whole genome cDNA microarrays have classified breast cancer as well as prostate cancer into molecular subtypes with distinct clinical and pathological characteristics. The present disclosure further extends the concept and the importance. Instead of only using statistical analyses, the selection of candidate gene markers in the present study was driven by a cancer stem cell (CSC)/embryonic stem cell (ESC) hypothesis, aiming at effectively identifying just a few most important ESC/CSC gene markers. This approach was proven to be effective since the most significant predictive gene markers identified in the present study were from the list of identified embryonic stem cell gene predictors (ESCGPs).
The inventors hypothesized that prostate cancer's biological aggressiveness and responsiveness to castration therapy are mainly determined by major gene expression patterns in prostate cancer stem cells (CSCs) (Visvader, Nature 2011, 469:314-22; Ratajczak et al, Differentiation 2011, 81:153-161; Lang et al, J Pathol 2009, 217:299-306). It was also hypothesized that genes that have important functions in embryonic stem cells (ESCs) can also have importance in prostate CSCs. Thus, direct measurement of the expression patterns of ESC related genes in prostate cancer cells would reflect the biological aggressiveness of the cancer and enable prediction of the effect of castration therapy as well as prediction of patient survival.
Based on this hypothesis the inventors have previously identified genes, i.e. embryonic stem cell gene predictors (ESCGPs) that have a consistently high or consistently low levels of expression in ESC lines (WO 2008/013492 A1). Briefly, the ESCGPs were identified by analyzing previously published datasets of whole genome cDNA microarray data derived from 5 human ESC lines and 115 human normal tissues from different organs by use of a simple one-class SAM, whereby the genes were ranked in order according to their degree of consistency in expression levels in the ESCs. This was based on the concept that genes with either consistently high or consistently low expression levels in all ESC lines may have significant functions in maintaining ESC status and their expression changes in different patterns can lead to differentiation toward different directions. These ESC genes may also have functions in maintaining different status of CSCs and thus different expression patterns of ESC genes in CSCs may classify tumors into different subtypes with different biological aggressiveness and sensibility to different types of treatments. Starting from this list of ESCGPs the present study identified some important prognostic and predictive gene markers for prostate cancer.
From the list of 641 ESCGPs identified in WO 2008/013492 A1 a subset of 33 ESCGPs were selected in the present study, as candidates that may enable classification of prostate cancers using fewer ESCGPs. The candidates were selected according to three criteria as described in Example 2A (see also
Furthermore 5 genes that were not from the ESCGP list were selected according to a fourth criterion; they were reported and known to be important in prostate cancer. The reported genes were used as controls to evaluate the importance of the ECSGP genes in relation to non-ESCGP genes in the classification of prostate cancer. Furthermore they could potentially be included in a molecular marker signature for use in prostate cancer classification.
Expression of the 33 selected ESCGPs and 5 reported genes in three different prostate cancer cell lines was investigated by RT-PCR (see Example 2B). Of the 33 genes 24 genes (19 ESCGPs and 5 reported genes) were identified that had different expression patterns in the less aggressive cell line LNCap compared to in the aggressive cell lines DU145 and PC3 (see
All patients in the present cohort had clinically significant prostate cancer and the majority (80%) of the patients were not treated by radical prostatectomy or full dose radiation therapy but only by castration therapy when the disease became advanced. Therefore, the survival data was not influenced by the cure effect of radical treatments, which when used at an early stage in some patients with biologically aggressive cancer could eliminate the cancer and thus the life threat. In the present cohort the time of follow up was 7-20 years and the majority (94.5%) of the patients were deceased, enabling a complete analysis of true overall survival time with minimal censored data. These characteristics ensured the discovery of new biomarkers for survival prediction and were unique as compared with most previous studies where PSA recurrence or progression free survival has been used as surrogate for overall and cancer specific survival.
In the present study both overall survival and cancer specific survival was used to evaluate the clinical value of prognostic biomarkers. Cancer specific survival is mainly determined by biological aggressiveness of cancer cells. However, the accuracy and importance of correlation between prognostic as well as predictive parameters (such as clinical parameters and/or expression of biomarkers) and cancer specific survival can be influenced by how cancer specific survival is defined, how the data is collected and how much data is censored due to other competing causes of mortality. On the other hand, overall survival is the survival data without any censoring by causes of death and including all causes of death. Therefore overall survival reflects not only biological aggressiveness of cancer cells but also many other factors such as competing disease or co-morbidities, complications as well as side effects of treatments, age and life expectancy. For prostate cancer patients overall survival may have more importance than cancer specific survival since most patients are diagnosed at old age and usually have other competing diseases such as cardiovascular diseases, diabetes mellitus or other malignant diseases (Daskivich et al, Cancer 2011, April 8. doi: 10.1002/cncr.26104. [Epub ahead of print]).
Ten molecular marker genes showed significant correlation with overall and/or cancer specific survival when analyzed by univariate analysis (see Table 1), and may be used for classification of prostate tumors, for prognosis prediction and also for making treatment decisions for patients depending on the classification of the patient's tumor. These were F3 (coagulation factor III), WNT5B (wingless-type MMTV integration site family, member 5B), VGLL3 (vestigial like 3 (Drosophila)), CTGF (connective tissue growth factor), IGFBP3 (insulin-like growth factor binding protein 3), c-MAF-a (long form of v-maf musculoaponeurotic fibrosarcoma oncogene homolog (avian)), c-MAF-b (short form of v-maf musculoaponeurotic fibrosarcoma oncogene homolog (avian)), AMACR (alpha-methylacyl-CoA racemase), MUC1 (mucin 1, cell surface associated) and EZH2 (enhancer of zeste homolog 2 (Drosophila)). Of these ten genes five (F3, WNT5B, CTGF, VGLL3 and IGFBP3) were ESCGPs identified from the list of genes with consistently high or low expression in embryonic stem cells. Two of the genes (c-MAF-a and c-MAF-b) were previously reported genes known to have important functions in myeloma.
Three of the significant genes (EZH2, AMACR and MUC1) are genes that have previously been reported in relation to prostate cancer. Several previous studies have identified biomarkers like AMACR, EZH2, MUC1 as well as AZGP1 and a sternness signature that are correlated to recurrence free survival after radical prostatectomy (Varambally et al, Nature 2002, 419:624-9; Rubin et al, JAMA 2002, 287:1662-70; Oon et al, Nat Rev Urol 2011, 8:131-8; Lapointe et al, Cancer Res 2007, 67:8504-10; Rubin et al, Cancer Epidemiol Biomarkers Prev
2005, 14:1424-32; Strawbridge et al, Biomark Insights 2008, 3:303-15; Glinsky et al, J Clin Oncol 2008, 2846-53; Glinsky et al, J Clin Invest 2005, 115:1503-21). The present results show that the expression level of MUC1, AMACR and EZH2 in prostate cancer FNA samples is indeed correlated to either cancer specific or overall survival. However, of the previously reported gene markers only the correlation by c-MAF-a was as strong as the correlation by the ESCGPs F3, IGFBP3 and VGLL3 identified in the present study.
The expression levels (reversely correlated to the delta Ct value) of all these significant genes but EZH2 showed positive correlation with survival times (Table 1, Hazard ratio>1). Only the expression level of EZH2 in the FNA samples was reversely correlated with patient survival. This negative correlation of EZH2 was in agreement with its documented role as an oncogene. The present results of EZH2, AMACR, IGFBP3 and c-MAF-a genes are in line with relevant results of previous studies (Varambally et al, Nature 2002, 419:624-9; Rubin et al, Cancer Epidemiol Biomarkers Prev 2005, 14:1424-32; Mehta et al, Cancer Res 2011, 71:5154-63; Li et al, Genes Chromosomes Cancer 1999, 24:175-82). IGFBP3 has well proven function in suppressing the metastatic process of prostate cancer (Mehta et al, Cancer Res 2011, 71:5154-63).
The positive correlation with survival by MUC1 and F3 was unexpected. F3 and MUC1 have documented functions in promoting cancer development (Strawbridge et al, Biomark Insights 2008, 3:303-15; Kasthuri et al, J Clin Oncol 2009, 27:4834-8). The positive correlation with survival may indicate that prostate cancer cells with high expression level of F3 and MUC1 are strongly androgen dependent and sensitive to castration treatment (Strawbridge et al, Biomark Insights 2008, 3:303-15; Kasthuri et al, J Clin Oncol 2009, 27:4834-8; Mitchell et al, Neoplasia 2002, 4:9-18; Brodin et al, Semin Thromb Hemost 2001, 37:87-94). There are some prognostic and predictive markers with similar dual aspects in other cancers, such as HER-2/neu/ERBB2 amplification in breast cancer, where breast cancer with HER-2/neu/ERBB2 amplification has aggressive biological as well as clinical features but shows response to Tratsuzumab (Herceptin) treatment with resulting prolonged survival.
The function of VGLL3 in prostate cancer is still unknown.
Multivariate analysis was further made in order to identify genes that show correlation to survival independent of all clinical parameters (see Example 3A). Four genes (F3, IGFBP3, CTGF and AMACR) showed correlation to both overall and cancer specific survival independent of all clinical parameters (
In order to study possible additive or synergic effects of multiple genes in the prediction of survival, different combinations of the ten significant genes in a series of unsupervised hierarchical clustering analyses were tested (see Example 3B and
The classification had strong correlation to and could be used for the prediction of a patient's overall and cancer specific survival (see
In addition the ability of the combination of genes IGFBP3 and F3 only (ESCGP signature 3) to classify tumor samples and predict survival was tested (see Example 3D). The tumor samples were first classified into three groups by use of unsupervised hierarchical clustering (
Most important marker genes found in this study showed correlation to both overall and cancer specific survivals. This was partly due to the possibility that prostate cancer or side effects of the treatments contributed also to deaths directly caused by other diseases. This could also partly be due to the fact that the ESCGP signature could be shared by both cancer stem cells and certain type of normal stem cells in the body. Thus, the ESCGP signature may have importance in the development of both cancer and other diseases. For instance, IGFBP3 has been identified with important suppressive functions in both cancer and diabetes (Yeap et al, Eur J Endocrinol 2011, 164:715-23; Mehta et al, Cancer Res 2011, 71:5451-63).
In a first aspect, the present invention provides a method of classifying prostate cancer in a subject, comprising:
In a preferred embodiment the expression level of the genes F3 and IGFBP3 and either of VGLL3 and c-MAF are determined in step a) and thus used for classification of the tumor. Preferably the expression level of F3, IGFBP3 and VGLL3 is determined.
These gene signatures have been shown particularly useful for the classification of prostate cancer tumors (
In one embodiment step a) thus further comprises determining a gene expression level of one or more of the genes VGLL3 and c-MAF, preferably VGLL3.
In a further embodiment step a) also comprises determining a gene expression level for one or more of the genes WNT5B and CTGF, EZH2, AMACR and MUC1.
In a second aspect, the present invention provides a method of classifying prostate cancer in a subject, comprising the steps of:
This second aspect of the invention is based on the herein recognized fact that expression of any of F3, IGFBP3, VGLL3, c-MAF, WNT5B and CTGF in samples from subjects having prostate cancer may serve as an indicator of disease status in said subject. The inventors have found that there is a positive correlation between the gene expression levels of any of said genes and survival. More particularly, the inventors of the present invention have found a correlation between a high level of expression of any of F3, IGFBP3, VGLL3, c-MAF, WNT5B and CTGF and longer survival, thus low risk tumors. On the other hand a low level of expression of either of said genes is correlated with shorter survival and thus high risk tumors.
In one embodiment of this second aspect the expression level of at least two, such as two, three or four of the genes F3, IGFBP3, VGLL3, c-MAF, WNT5B and CTGF is determined in step a) of the method according to the present invention and thus used for classification of the tumor.
In a further embodiment the expression level of all of the genes F3, IGFBP3, VGLL3, c-MAF, WNT5B and CTGF is determined in step a) of the method according to the present invention and thus used for classification of the tumor.
In still another embodiment the expression level is also, that is in addition to any of the combinations above, determined for at least one of the genes EZH2, AMACR and MUC1 and thus used in the classification.
Whether the expression level of one of said genes in a patient with prostate cancer is high or low can be determined by comparing the gene expression level in a sample from the patient with a reference gene expression value of the same gene(s) in a reference patient, or group of reference patients, known to have a high risk or low risk tumor respectively. If the expression level of the selected gene(s) in the patient sample is as high as or higher than the expression level of the same gene in a reference patient known to have a low risk tumor the tumor of the subject may be classified as being low risk. If the expression level of the selected gene(s) in the patient sample is as low as or lower than the expression level of the same gene in a reference patient known to have a high risk tumor the tumor of the subject may be classified as being high risk. When a group of reference patients is used for the comparison, the medium or median expression level of the selected gene(s) in the group may be used as the reference gene expression value.
By matching the gene expression level of the selected gene with the reference gene expression of a reference patient is meant, when the gene expression level is determined for one gene, that when the expression level of the selected gene is as high as or higher than the reference gene expression in a reference patient known to have a low risk tumor the gene expression level matches that reference gene expression. Likewise, when the expression level of the selected gene is as low as or lower than the reference gene expression in a reference patient known to have a high risk tumor the gene expression level matches that reference gene expression.
By matching the gene expression level of the selected gene with the reference gene expression of a reference patient is meant, when the gene expression level is determined for two or more genes, that the overall gene expression pattern of the two or more selected genes must match with the overall reference gene expression pattern of the two or more selected genes in a reference patient. Thus the expression of both or all of the selected genes, as evaluated one by one, need not completely match the reference gene expression of the selected genes one by one. Rather a very high level of gene expression for one of the genes may compensate for a lower level of the other gene(s) and the expression pattern would still be considered matching. By gene expression pattern is meant the gene expression level for the genes in a selection of two or more genes.
Matching of gene expression profiles obtained from the subject and the reference patient respectively may for instance be made using hierarchical clustering using the gene expression data from both the subject and the reference samples, by way of methods that are known in the art (see e.g. Eisen et al, Proc Natl Acad Sci USA 1998, 95:14863-8). Clustering methods are suitable for evaluating trends in large data sets. Unsupervised clustering like hierarchical clustering is advantageously used to detect groups or classes in data sets that would not easily be recognized by just browsing the data. If the patient whose tumor is to be classified is clustered or grouped together with reference patients that are known have a low risk tumor then the tumor of the patient is also classified as a low risk tumor. If the patient whose tumor is to be classified is clustered or grouped together with reference patients that are known have a high risk tumor then the tumor of the patient is also classified as a high risk tumor.
By a high risk tumor is meant that the tumor subtype, as determined by use of a group of patients with known tumor subtype and known survival, is associated with shorter overall and/or cancer specific survival time than a low risk tumor. The subtype may for instance be defined as a tumor subtype with certain clinical parameters or with certain expression of certain genes. When determining whether there is a significant difference in survival time between patients with known subtypes and known survival times one may use calculation of hazard ratios which is well known in the art (Cox D R, J Royal Statist Soc B 1972, 34:187-220). A hazard in a group is the rate at which events, such as death, happens. The hazard in one group is assumed to be a constant proportion of the hazard in the other group. This proportion is the hazard ratio. Thus, if the hazard ratio is, with significance, higher or lower than one there is a higher risk for one group over the other.
The classification of the tumor may also include more classes than high risk and low risk such as one or more intermediate risk group(s).
The sample from the subject may be a tumor sample, such as a tumor sample obtained by fine needle aspiration (FNA), needle biopsy or by surgery. Alternatively the sample may be a blood sample, plasma, serum, cerebral fluid, urine, semen, exudate or a stool sample obtained from the subject. In particular, the gene expression level for IGFBP3 and F3 may advantageously be determined by analysis of a blood sample.
In one embodiment the gene expression level of the selected genes is determined by quantifying the amount of RNA or mRNA expressed from the genes. The amount of RNA or mRNA may for instance be determined by use of a method selected from microarray technology, Northern blotting and quantitative PCR (qPCR), such as real time quantitative PCR (qrt-PCR), optionally multiplex PCR, or any other method for measurement of gene expression known in the art.
For instance, the inventors have in the present study developed a simple multiplex quantitative PCR (qPCR) method to measure expression levels of several selected marker genes in prostate fine needle aspiration (FNA) samples. The developed method may also be used to measure expression levels in any tumor or blood sample taken from the patient.
One important technical advantage of this approach is that although the marker genes of the present invention are identified through a stem cell approach and are believed to be important for cancer stem cell function one does not need to directly isolate the CSCs from the tumor samples. The simple and robust multiplex qPCR method established in the present study can be directly applied to analyze fresh samples from routine needle biopsy or aspiration cytology for the prediction of survival and effect of castration therapy at the time of diagnosis. All samples analyzed in the present study were fresh-frozen cytological cell spreads that would ensure isolation of pure cancer cell RNAs of high quality for qPCR analysis. However, the RNA isolation was not successful in some cases due to too few cells on the glass slides of FNA cytology spreads. The problem can be easily solved in the future clinical application by directly using fresh FNA cell suspensions or microdissected tumor samples from the needle biopsies for RNA isolation.
Since the marker genes of the present invention (F3, IGFBP3, VGLL3, c-MAF, WNT5B and CTGF) encode proteins it is also possible to use immunochemistry or other protein analytical methods to measure their protein expression as an estimate or function of their gene expression. Thus in one embodiment of the present invention the gene expression level may be indirectly determined by measuring the amount of protein encoded by said genes. The amount of protein may for instance be determined by use of methods such as immunohistochemistry, Western blotting, enzyme immunoassays such as ELISA, RIA and mass spectrometry, as well as other methods for protein detection known in the art.
The skilled person will recognize that the usefulness of the present invention is not limited to the quantification of gene expression of any particular variant of the marker genes of the present invention. As non-limiting examples, the marker genes may have coding sequences and amino acid sequences as specified in Table 6. In some embodiments they have cDNA sequences or amino acid sequences that are at least 85% identical or similar to the listed sequences, such as at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or at least 99% identical or similar to the sequences listed in Table 6.
The term “% identity”, as used throughout the specification, is calculated as follows. The query sequence is aligned to the target sequence using the CLUSTAL W algorithm (Thompson, J. D., Higgins, D. G. and Gibson, T. J., Nucleic Acids Research, 22: 4673-4680 (1994)). A comparison is made over the window corresponding to the shortest of the aligned sequences. The shortest of the aligned sequences may in some instances be the target sequence. In other instances, the query sequence may constitute the shortest of the aligned sequences. The amino acid residues at each position are compared, and the percentage of positions in the query sequence that have identical correspondences in the target sequence is reported as % identity.
The term “% similarity”, as used throughout the specification, is calculated in the following way. Sequence alignment and comparison are basically performed as described in relation to the % identity calculation. “Similarity” should however be interpreted as follows. Two amino acid residues are considered similar if they belong to the same group of amino acid residues. Non-limiting examples of groups of amino acid residues are the hydrophobic group, comprising the amino acid residues Ala, Val, Phe, Pro, Leu, Ile, Trp, Met and Cys; the basic group, comprising the amino acid residues Lys, Arg and His; the acidic group, comprising the amino acid residues Glu and Asp; the hydrophilic group, comprising the uncharged amino acid residues Gln, Asn, Ser, Thr and Tyr; and the natural group, comprising the amino acid residue Gly. Thus, the amino acid residues at each position are compared, and the percentage of positions in the query sequence that have similar correspondences in the target sequence is reported as % similarity.
The method for classifying a tumor in a subject having prostate cancer according to the present invention may have many benefits. For example, as in one embodiment of the invention, it may be used to predict survival for said subject. For a subject with a tumor that is classified to be a low risk tumor it is indicated that the subject has a good prognosis, while for a subject with a tumor that is classified to be a high risk tumor it is indicated that the subject has a poor prognosis.
A poor prognosis for a subject may mean that the subject has a decreased likelihood of survival or decrease in time of survival compared to a subject that has been predicted to have a good prognosis. A poor prognosis may also mean that the patient has an increased risk of recurrence or metastasis as compared to a patient with a good prognosis. For example, the likelihood of five-year survival for a patient with a low risk tumor may be 90% or lower, such as 85%, 80%, 75%, 70%, 60% or lower, while the likelihood of five-year survival in the high risk group may be 50% or lower, such as 45%, 40%, 30%, 20%, 10% or lower. The median length of survival for patients with low risk tumors may likewise be 6 years or longer, such as 7 years, 8 years, 9 years, 10 years or longer, while the median length of survival in patients with high risk tumors may be 5 years or shorter, such as 4 years, 3 years, 2 years, 1 year or shorter.
In one embodiment of the invention the classification of a tumor may be used to improve survival prediction using clinical parameters. The inventors have for example shown (Example 3C) that when subtype classification by use of Signature 1 (VGLL3, IGFBP3 and F3) is added to conventional prediction models using only clinical parameters the accuracy of prediction is significantly improved.
In one aspect the invention provides a method for taking a decision on future treatment for the patient, the decision being dependent on the classification according to the invention. Patients that have a tumor that has been classified as being a high risk tumor need more radical or curative treatments than patients with low risk tumors, and also at an earlier stage. Radical or curative treatments include treatment regimes selected from prostatectomy, radiation, chemotherapy, castration or a combination thereof. Patients with tumors that have been classified as low risk tumors need less or no radical or curative treatment, but can be assigned to watchful-waiting or active surveillance. In certain embodiments of the invention patients with localized cancer of high risk or intermediate risk tumor subtype need radical or curative treatments without delay, while patients with localized cancer of low risk tumor subtype can be safely assigned to watchful waiting with minimal anxiety because castration therapy can be still a guarantee of long time survival in case of disease progression. For patients with advanced cancer at diagnosis, those of low risk subtype can get most benefit from castration therapy or anti-androgen therapy whereas patients of high risk and intermediate risk subtype may need to be treated by chemotherapy or other new therapies early.
In one aspect the invention further provides a method of treating the subject that has been diagnosed with prostate cancer, and whose tumor has been classified according to the invention, in accordance with the treatment decision made as above.
In one aspect the invention provides use of any of the genes IGFBP3, F3, VGLL3, c-MAF, WNT5B and/or CTGF or the proteins encoded therefrom as prognostic marker(s) for prostate cancer. In various embodiments of this aspect the invention provides use of a combination of two, three or more of the genes IGFBP3, F3, VGLL3, c-MAF, WNT5B and/or CTGF or the proteins encoded therefrom as prognostic markers for prostate cancer. One particularly useful embodiment provides use of a combination of the genes IGFBP3 and F3 and, optionally, either of VGLL3 and c-MAF, or the proteins encoded therefrom as prognostic markers for prostate cancer.
In one aspect the invention provides a solid support or a kit for classifying a tumor in a subject diagnosed with prostate cancer, comprising nucleic acid probes or antibodies that are useful for determining gene expression and are specific for a combination of at least two of the genes IGFBP3, F3, VGLL3, c-MAF, WNT5B and CTGF. In one embodiment thereof, said solid support or kit comprises nucleic acid probes or antibodies that are specific for IGFBP3 and F3. In another embodiment the solid support or kit comprises nucleic acid probes or antibodies that are specific for IGFBP3 and F3 and either or both of VGLL3 and c-MAF. In still another embodiment the solid support or kit further comprises nucleic acid probes or antibodies that are specific for EZH2, AMACR and MUC1.
The solid support may be an array, such as a cDNA microarray, a polynucleotide array or a protein array.
The nucleic acid probes for any of the kit embodiments may for example be selected from the sequences disclosed in Table 7. Such kit is particularly useful for determination of gene expression levels using multiplex PCR, e.g. multiplex quantitative PCR.
The kit may also comprise further reagents that are necessary for the measurement of gene expression level, such as secondary labeled probes or affinity ligands for detecting and/or quantifying bound or amplified nucleic acids or antibodies, depending on the selected method. Such labels may also be directly attached or linked to the nucleic acid probes or antibodies.
The kit may further comprise various auxiliary substances to enable the kit to be used easily and efficiently, e.g. solvents, wash buffers etc. In addition the kit may also advantageously comprise reference samples or information on reference gene expression level values obtained by use of the same method from patients with known high risk or low risk tumors.
General Methods
Bioinformatics Analysis
Bioinformatics analysis for identification of embryonic stem cell gene predictors (ESCGPs) has been described previously (WO 2008/013492 A1). Briefly, previously published cDNA microarray gene expression datasets were retrieved from the Stanford Microarray Database (SMD, http://smd.stanford.edu/). The criterions used for data retrieving were as following:
Gene/spot selection: all genes or clones on arrays were selected, control spots and empty spots were not included.
Data Collapse and Retrieval: row data were retrieved and averaged by SUID; UID column contains NAME.
Data Retrieved: Log(base2) of R/G Normalized Ratio (Mean).
Selected Data Filters: Spot is not flagged by experimenter.
Data filters for GENEPIX result sets: Channel 1 Mean Intensity/Median Background Intensity>1.5 AND Channel 2 Normalized (Mean Intensity/Median Background Intensity)>1.5.
Cluster program (version 3.0) was used to carry out unsupervised hierarchical average linkage clustering and TreeView program to visualize the cluster results (Eisen et al, Proc Natl Acad Sci USA 1998, 95:14863-8). SAM (significant analysis of microarrays) was carried out as previously described (Tusher et al, Proc Natl Acad Sci USA 2001, 98:5116-21).
Data Centering of Retrieved cDNA Microarray Dataset: The cDNA microarray data of 5 human ESC lines (Sperger et al, Proc Natl Acad Sci USA 2003, 100:13350-5) and 115 human normal tissues from different organs (Shyamsundar et al, Genome Biol 2005, 6:R22) were retrieved from the SMD according to parameters described in the above. The dataset was divided into subsets by different array batches. Genes were centered within each array batch by using the gene centering function of the Cluster program. The subsets were combined again and arrays were centered by using the array centering function of the Cluster program. After centering the dataset was saved and converted into Excel form.
Prostate Cancer Cell Lines
Three prostate cancer cell lines LNCaP, DU145 and PC3 were purchased from the American Type Culture Collection (ATCC). Cell culture was carried out with medium and methods according to the instruction by ATCC. LNCaP, DU145 and PC3 Cells are maintained by Iscove's Modified Dulbecco's Medium (IMDM, Cat No. 21980-032, Invitrogen) supplemented by 10% Fetal Bovine Serum (Cat No. 10082-147, Invitrogen) and 50 unit/ml and 50 ug/ml Penicillin/Streptomycin (Cat No. 15140-163, Invitrogen).
FNA Samples
Prostate FNA (fine needle aspiration) samples were taken by routine procedure for cytology diagnosis at the Department of Clinical Cytology and Pathology, Karolinska Hospital, Stockholm, Sweden. FNA samples were obtained from 241 patients at the time of diagnosis before any treatments. At least one fresh cytology spread from each patient was Giemsa stained for clinical cytology diagnosis. Remaining duplicate fresh spreads were transferred to deep freezer and had been kept fresh frozen at 80° C. until the isolation of RNA samples. Most FNA cytology spreads with prostate cancer diagnosis were estimated to contain over 80% of tumor cells due to the well known selecting effect that the aspiration sampling process can enrich cancer cells due to their decreased cell adhesion. Of the 241 patients, isolation of RNA with good quality was successful in samples from 193 patients. Of those 189 were diagnosed with prostate cancer while 4 patients did not have prostate cancer.
Clinical Characteristics of the Cohort
In total freshly frozen FNA samples from 189 prostate cancer patients were analyzed in the present study. These 189 prostate cancer patients were diagnosed during years 1986-2001. All the 189 patients had clinical symptoms which led to the diagnosis of prostate cancer. Under oncologist supervision an internship doctor collected relevant clinical data such as age at diagnosis, date of diagnosis, cytology and biopsy diagnosis, serum PSA at diagnosis, clinical stage, primary treatment, etc. Table 8 presents details about clinical characteristics of these 189 patients.
Data for date of diagnosis, date of death and causes of death for all patients were first obtained from regional as well national registries and then verified by available original medical journals. The date for data censoring was the 31 Dec. 2008. By this time, of the 189 patients 22 were still alive, 163 were deceased and 4 were without data in the registries. Prostate cancer specific death was defined as that the primary or secondary cause of death was prostate cancer or metastases. Death due to other causes was defined as the primary and secondary causes of death were not prostate cancer or metastases. These cases included even patients who died of diseases or conditions that could become worse due to prostate cancer or related to side effects and complications of treatments.
All the 189 patients had clinical symptoms which led to digital rectal examination, PSA test and subsequent prostate FNA. Castration therapy was the only primary treatment for most patients (77.9%) when the disease became advanced.
RNA Isolation
AliPrep DNA/RNA Mini Kit (Cat No. 80204, QIAGEN) was used for total RNA isolation in prostate cancer cell lines. RNAqueous®-Micro Kit (Cat No. 1931, Ambion) for isolation of total RNA less than 100 ng was used to isolate total RNAs from freshly frozen FNA samples from prostate cancer patients. RNA quantity and quality were controlled by using Agilent RNA 6000 Nano Kit (Cat No. 5067-1511, Agilent) on a 2100 RNA Bioanalyzer (Agilent). RNA samples with RNA integrity number (RIN) larger than 7 were considered as qualified. In the present study, qualified total RNA was isolated from 193 of the 241 FNA samples for further cDNA synthesis and qPCR experiments.
RT-PCR
For reverse transcription (RT) reactions, cDNA synthesis for PCR (polymerase chain reaction) was carried out by using a Cloned AMV First-Strand cDNA Synthesis Kit (Cat No. 12328-032, Invitrogen) according to the manufacturer's instruction. Maximally 2 ug total RNA was used for RT in 20 ul reaction volume. The expression patterns of 33 ESCGPs and 5 control genes in prostate cancer cell lines were validated by RT-PCR using gene specific primer pairs (
4-Plex Real Time qPCR
First-strand cDNA synthesis for quantitative PCR (qPCR) was run using a QuantiTect® Reverse Transcription Kit (Cat No. 205311, QIAGEN). Up to 1 ug total RNA was used for each qPCR in 20 ul reaction volume. The reaction was run on an ABI 7500 real time cycler that could in real time simultaneously monitor the densities of four different fluorescent dyes (4-plex). None passive reference was selected in this four-dye combination. The condition for 4-plex qPCR was at 50° C. for 2 minutes in 1 cycle; at 94° C. for 10 minutes in 1 cycle; at 94° C. for 1 minutes in 40 cycles and at 60° C. for 1.5 minutes in 1 cycle. Fixed baseline start value and end value were chosen for Ct value analysis (Schmittgen and Livak, Nat Protoc 2008, 3:1101-8; Wittwer et al, Methods 2001, 25:430-42).
Optimization of 4-Plex Real Time qPCR
A 4-plex qPCR contains four pairs of gene specific primers and four gene specific Taqman probes each of which was dual-labeled with a fluorophore on the 5′ end and a quencher on the 3′ end. In our study, Cy5, FAM, Texas Red and VIC were used for the 5′ end labeling while BHQ-3, BHQ-1, BHQ-2 and TAMRA were used as the 3′ quenchers. The four different combinations of the fluorophore-quencher pair enabled specific detection of PCR products of the 4 different genes. In total, For 19 ESCGPs and 5 control genes, 45 predicted 4-plex probes and 24 pairs of primers were designed by Beacon Designer 7.0 software (Primer Biosoft). Sequence information of probes and primers for the genes of the present invention is presented in Table 7.
To validate whether 4-plex qPCR has the same specificity and efficiency with single probe qPCR, cDNA standard curve method was used. cDNAs derived from total RNAs purified from LNCap, DU145 and PC3 cells were diluted to a series of concentrations at 10 pg, 100 pg, 1000 pg, 10000 pg, 100000 pg were used as templates for both single probe qPCR and 4-plex qPCR respectively. Standard curves are made based on the Ct value of each probe and the amount of cDNAs. The values of slope and r of cDNA standard curves derived from single probe qPCR and 4-plex qPCR of the same genes were compared. Optimization of concentrations of probes and primer pairs was carried out until there was no significant difference in these values between single and 4-plex qPCR. The results showed that 0.2 uM probes and 0.2 uM primer pairs were the best concentrations for 4-plex qPCR. Validation results of 4-plex qPCR are presented in
Normalization and Centering of qPCR Result Ct Value
Ct (cycle threshold) is a measure of the number of PCR cycles (in real-time PCRs) needed to obtain a fluorescent signal or enough PCR products. In the present study, Ct value of a gene in a sample after real time PCR was generated by using 7500 software (version 2.0.5, ABI). In order to normalize the Ct values of each gene, delta Ct value was calculated according to an equation ΔCt=CtgeneX−CtGAPDH where CtgeneX was the Ct value of the gene to be analyzed and CtGAPDH was the Ct value of the housekeeping gene GAPDH (glyceraldehyde-3-phosphate dehydrogenase) (Schmittgen and Livak, Nat Protoc 2008, 3:1101-8; Wittwer et al, Methods 2001, 25:430-42). Thus, the expression level of each gene in a sample was normalized by the expression level of GAPDH. The ΔCt was reversely correlated with the gene expression level. Each panel of 4-plex qPCR contains one specific GAPDH probe respectively. Samples with weak signals were excluded from analysis (Ct value of GAPDH>28). Samples with weak signals of genes to be analyzed, their Ct values were set as 40 (set as the maximal value of Ct). Delta Ct values of genes in all samples were centered by using the gene median center function of a Cluster program (version 3.0) (Eisen et al, Proc Natl Acad Sci USA 1998, 95:14863-8). The centered delta Ct value was used for statistical analyses.
Statistical Analysis of Survival Correlation
Overall survival and prostate cancer specific survival were used as the endpoints respectively in survival analysis for the correlation with molecular and clinical parameters. Survival time was defined as the time from the date of diagnosis to the date of death and was used as continuous variable. For simplified interpretation, long, intermediate or short survival was defined as survival time>8, 5-8 or <5 years respectively. For patients treated primarily only by castration therapy the leading time before the treatment was defined as the time from the date of diagnosis to the date of start of castration treatment and was used as continuous variable. The centered delta Ct value of each gene, age at diagnosis and serum PSA value at diagnosis were used as continuous variables. By unsupervised hierarchical clustering analysis samples were classified into three groups or subtypes and the grouping was used as non-continuous variable. PSA was also analyzed as non-continuous variable by two categories ≤50 ng/ml or >50 ng/ml. The WHO tumor grade was integrated into two categories: well-moderate differentiated or poorly differentiated. The clinical stage was integrated into two categories: advanced (any T≥T3 or N1 or M1 or PSA≥100 ng/ml) or localized (T<T3 and N0 and M0 and PSA<100 ng/ml). Univariate as well as multivariate analyses of Cox proportional hazard ratio and Cox regression were performed by Stata (Version 10.1, StataCorp LP) statistics software. Kaplan-Meier analysis as well as statistic box plots were carried out by using JMP® statistics software (version 8.0.1, SAS Institute Inc).
Study Set Up
The study was carried out in three steps:
This resulted in identification of a subset of gene markers that show a significant correlation to either overall or cancer specific survival.
An ESCGP signature for classification of various types of cancers was identified as disclosed in patent document WO 2008/013492 A1. Briefly, previously published datasets of whole genome cDNA microarray data derived from 5 human ESC lines and 115 human normal tissues from different organs were retrieved from the Stanford Microarray Database (SMD) according to parameters described above. Data centering of the retrieved datasets was also carried out as described above. Data from the normal tissues were used to aid the data centering. After centering the sub-dataset of the ESC lines was isolated from the whole dataset. A one class SAM was carried out by using only this ESC line dataset, by which all genes were ranked according to the consistency of their expression levels across the 5 ESC lines. By using a q-value 0.05 as cut-off the analysis identified 328 genes with consistently high and 313 genes consistently low expression levels in the ESCs. The 641 genes were named as embryonic stem cell gene predictors (ESCGPs).
From the list of 641 ESCGPs a subset of 33 ESCGPs as well as 5 control genes were selected as candidates that may enable classification of prostate cancers using fewer ESCGPs. The candidates were selected according to four criteria (see
Expression of the 33 selected ESCGPs and 5 control genes in three different prostate cancer cell lines were validated by RT-PCR using gene specific primer pairs (see
Expression of the 24 genes (25 gene markers) was analyzed in fine needle aspiration (FNA) samples from 189 prostate cancer patients by use of multiplex qPCR, and then analyzed for correlation with survival data. Clinical characteristics of the patient cohort as well as the statistical analysis is described above.
All candidate genes could not be analyzed in every FNA sample due to small amount of total RNA from most FNA samples. To compromise the limitation, the cohort of 189 patients was divided into three sets according to the experiment time order. The three sets contained samples from 36, 65 and 88 patients respectively (Table 8). Only genes that showed significant correlation with survival in the first subset were included together with new candidate genes in the subsequent subset. Survival analysis was carried out in each of the three subsets as well as in the final complete cohort (Table 1,
Analysis of correlation with survival was carried out for both clinical parameters known for the patients and for gene expression of the selected candidate genes. In univariate analysis all clinical parameters showed significant correlation with both overall and cancer specific survival (Table 1). Ten of the 25 gene markers, F3 (coagulation factor III), WNT5B (wingless-type MMTV integration site family, member 5B), VGLL3 (vestigial like 3 (Drosophila)), CTGF (connective tissue growth factor), IGFBP3 (insulin-like growth factor binding protein 3), c-MAF-a (long form of v-maf musculoaponeurotic fibrosarcoma oncogene homolog (avian)), c-MAF-b (short form of v-maf musculoaponeurotic fibrosarcoma oncogene homolog (avian)), AMACR (alpha-methylacyl-CoA racemase), MUC1 (mucin 1, cell surface associated) and EZH2 (enhancer of zeste homolog 2 (Drosophila)) showed significant correlation with either overall and/or cancer specific survival (Table 1). A p-value<0.05 is considered significant throughout the study. The expression levels (reversely correlated to the delta Ct value) of all these significant genes but EZH2 showed positive correlation with survival times (value<1 in Table 1).
Each of the ten gene markers with significant correlation with survival in univariate analysis was analyzed together with clinical parameters including age at diagnosis, two-category PSA, tumor grade and clinical stage in multivariate analysis (
In order to study possible additive or synergic effects of multiple genes in the prediction of survival, the inventors tested different combinations of the ten significant genes in a series of unsupervised hierarchical clustering analyses using the data of patients in the first set (training set). Two signatures could in a similar manner classify tumors into three subgroups or subtypes with significant difference in overall and cancer specific survival (
The ESCGP Signature 1 (VGLL3, IGFBP3 and F3) showed better results than the ESCGP Signature 2 (c-MAF-a, IGFBP3 and F3) (Tables 2 and 3). Of the 189 patients, 87 had data for both all clinical parameters and for the subtype classification by Signature 1. Multivariate analysis for overall and cancer specific survival showed that the subtype classification by Signature 1 was the most significant parameter and independent of age, PSA level, tumor grade and clinical stage (Table 2).
Median overall survival was 2.60 years in the high risk, 3.85 years in the intermediate risk and 7.98 years in the low risk subtype (
Interestingly, median survival time of unspecific deaths was 3.54 years in the high risk, 3.70 years in the intermediate risk and 7.98 years in the low risk subtype (
Kaplan-Meier curves further presented obvious survival difference between the three subtypes classified by the tumor ESCGP Signature 1. Overall survival rate of high risk (Group 1), intermediate risk (Group 2) and low risk (Group 3) subtype was 20%, 40% and 80% at 5 years, and 10.3%, 25.0% and 64.4% at 8 years respectively (
The survival difference between the high risk and the low risk subtype was much more impressive than the results by any clinical parameters, and was still seen within each patient group or became further more obvious within the same patient group defined by PSA, clinical stage, tumor grade or age (
Parametric model was used for survival prediction to estimate how much the subtype classification by the signature of VGLL3, IGFBP3 and F3 (Signature 1) could improve the prediction by using all clinical parameters (
Out of 189 patients, 95 had data available for the evaluation of ESCGP signature 3 (IGFBP3 and F3). Three tumor subtypes (Group 1, Group 2 and Group 3) were classified by unsupervised hierarchical clustering method using the median-centered delta Ct values of the two genes (F3 and IGFBP3) measured in the FNA samples. The results were visualized by using the Treeview software (
| Number | Date | Country | Kind |
|---|---|---|---|
| 1150982-5 | Oct 2011 | SE | national |
This application is a continuation application of U.S. patent application Ser. No. 15/725,678, filed Oct. 5, 2017, which is a continuation application of U.S. patent application Ser. No. 14/352,107, filed Apr. 16, 2014 (U.S. Pat. No. 9,790,555), which is a national phase application of International Application No. PCT/EP2012/071077 filed on Oct. 24, 2012, which claims priority to Swedish Application No. 1150982-5 filed on Oct. 24, 2011. The contents of these earlier filed applications are hereby incorporated by reference in their entirety.
| Number | Name | Date | Kind |
|---|---|---|---|
| 9790555 | Li | Oct 2017 | B2 |
| 20010053519 | Fodor et al. | Dec 2001 | A1 |
| 20020038464 | Charron et al. | Mar 2002 | A1 |
| 20030054419 | Slawin et al. | Mar 2003 | A1 |
| 20090298082 | Klee et al. | Dec 2009 | A1 |
| 20100056445 | Sharma et al. | Mar 2010 | A1 |
| 20100233691 | Bankaitis-Davis et al. | Sep 2010 | A1 |
| 20110166030 | Wang et al. | Jul 2011 | A1 |
| 20180080088 | Li et al. | Mar 2018 | A1 |
| Number | Date | Country |
|---|---|---|
| 2011-517283 | Jun 2011 | JP |
| WO2009021338 | Feb 2009 | NO |
| WO2001071355 | Sep 2001 | WO |
| WO2006028867 | Mar 2006 | WO |
| WO2006135886 | Dec 2006 | WO |
| WO2008013492 | Jan 2008 | WO |
| WO-2008013492 | Jan 2008 | WO |
| WO2008079269 | Jul 2008 | WO |
| WO2009105154 | Aug 2009 | WO |
| WO2009108860 | Sep 2009 | WO |
| WO2009143603 | Dec 2009 | WO |
| WO2010006048 | Jan 2010 | WO |
| WO2010101888 | Sep 2010 | WO |
| WO2013060739 | May 2013 | WO |
| Entry |
|---|
| Prueitt et al. (Prostate 68:1152-1164, 2008) (Year: 2008). |
| Blumke et al. (The Journal of Urology, vol. 175, No. 4, Supplement, Sunday May 21, 2006, Abstract #249, p. 83) (Year: 2006). |
| Keys et al. Chapter 6 in Sridar V. Chittur (ed.), Microarray Methods for Drug Discovery, Methods in Molecular Biology, vol. 632, DOI 10.1007/978-1-60761-663-4_6, © Humana Press, a part of Springer Science+Business Media, LLC 2010. (Year: 2010). |
| NCBI Geo record GSE7055. Obtained from Geo Gene Expression Omnibus database. 2 pages. (Year: 2008). |
| Definition of Tumor. NCI dictionary of cancer terms. https://www.cancer.gov/publications/dictionaries/cancer-terms/def/tumor Accessed Jun. 19, 2023. one page. (Year: 2023). |
| Agilent-014850 Whole Genome Microarray, 4X44K, as disclosed in NCBI Platform GPL4133 (2006). |
| Akashi, T. et al: “Tissue factor expression and prognosis in patients with metastatic prostate cancer”. Uroloov. (2003) 62(6):1078-1082. |
| Chandran, U., et al. (2007), “Gene expression profiles of prostate cancer reveal involvement of multiple molecular pathways in the metastatic process”, BMC Cancer, 7(64): 21 pages. |
| Flaig, T., et at. (2007), “Conference Report and Review: Current Status of Biomarkers Potentially Associated With Prostate Cancer Outcomes”, The Journal of Uroloov, 177: 1229- 1237. |
| Iscove et al. Representation is faithfully preserved in global eDNA amplified exponentially from sub-picogram quantities of mRNA, Nature Biotechnology, (2002), p. 940-943. |
| Jing, M., et al. (2009), “Prediagnostic plasma levels of insulin-like growth factor (IGF)-1, IGF-binding protein (IGFBP)-3, and C-peptide predict long-term prostate cancer survival in men without apparent prostate cancer at baseline”, Proceedings of the American Association for Cancer Research, Epidemioloov 9, 50: 733. |
| Kurashige, Y., et al. (2008), “Profiling of differentially expressed genes in procine epithelial cells derived from periodontal ligament and gingiva by DNA microarray”, Archives of Oral Biology, 53: 437-442. |
| Lapointe, J., et al. (2004), “Gene expression profiling identifies clinically relevant subtypes of prostate cancer”, PNAS, 101(3): 811-816. |
| Mehta, H et al: “OR6,34 IGFBP-3 is a metastasis suppression gene for prostate cancer”. Growth Hormone and IGF Research (2010) 20: S15. |
| Nakagawa, et al: A Tissue Biomarker Panel Predicting Systemic Progression after PSA Recurrence Post-Definitive Prostate Cancer Therapy II. Plos One. (2008).vol. 3(5): e2318. |
| NEB catalog (1998/1999), pp. 121, 284. |
| Papadimitrakopoulou, V., et al. (2006), “The prognostic role of loss of insulin-like growth factorbinding protein-3 expression in head and neck carcinogenesis”, Cancer Letters, 239: 136-143. |
| Pradervand, S., et al. (2008), “Affymetrix Whole-Transcript Human Gene 1.0 ST array is highly concordant with standard 3′ expression arrays”, BioTechniques, 44(6): 759-762. |
| Strijbos, M.H. et al: “Circulating endothelial cells, circulating tumour cells, tissue factor, endothelin-1 and overall survival in prostate cancer patients treated with docetaxel”. European Journal of Cancer. (2010) 46(11): 2027-2035. |
| International Search Report and Written Opinion mailed on Oct. 1, 2013 by the International Searching Authority for International Patent Application No. PCT/EP2012/071077, which was filed on Oct. 24, 2012 and published as WO 2013/060739 on May 2, 2013 (Applicant—Chundsell Medicals AB) (11 pages). |
| International Preliminary Report on Patentability issued on Apr. 29, 2014 by the International Searching Authority for International Patent Application No. PCT/EP2012/071077, which was filed on Oct. 24, 2012 and published as WO 2013/060739 on May 2, 2013 (Applicant—Chundsell Medicals AB) (7 pages). |
| Preliminary Amendment mailed on Apr. 16, 2014 by the USPTO for U.S. Appl. No. 14/352,107, which was filed on Apr. 16, 2014 and granted as U.S. Pat. No. 9,790,555 on Oct. 17, 2017 (Inventor—Chunde Li, et al.) (10 pages). |
| Restriction Requirement issued on Aug. 18, 2016 by the USPTO for U.S. Appl. No. 14/352,107, which was filed on Apr. 16, 2014 and granted as U.S. Pat. No. 9,790,555 on Oct. 17, 2017 (Inventor—Chunde Li, et al.) (9 pages). |
| Response to Restriction Requirement mailed on Oct. 11, 2016 by the USPTO for U.S. Appl. No. 14/352,107, which was filed on Apr. 16, 2014 and granted as U.S. Pat. No. 9,790,555 on Oct. 17, 2017 (Inventor—Chunde Li, et al.) (7 pages). |
| Non Final Rejection issued on Jan. 20, 2017 by the USPTO for U.S. Appl. No. 14/352,107, which was filed on Apr. 16, 2014 and granted as U.S. Pat. No. 9,790,555 on Oct. 17, 2017 (Inventor—Chunde Li, et al.) (14 pages). |
| Response to Non Final Rejection mailed on May 23, 2017 by the USPTO for U.S. Appl. No. 14/352,107, which was filed on Apr. 16, 2014 and granted as U.S. Pat. No. 9,790,555 on Oct. 17, 2017 (Inventor—Chunde Li, et al.) (12 pages). |
| Notice of Allowance issued on Jun. 9, 2017 by the USPTO for U.S. Appl. No. 14/352,107, which was filed on Apr. 16, 2014 and granted as U.S. Pat. No. 9,790,555 on Oct. 17, 2017 (Inventor—Chunde Li, et al.) (9 pages). |
| Notice of Allowance issued on Jun. 28, 2017 by the USPTO for U.S. Appl. No. 14/352,107, which was filed on Apr. 16, 2014 and granted as U.S. Pat. No. 9,790,555 on Oct. 17, 2017 (Inventor—Chunde Li, et al.) (2 pages). |
| Amendment/Req. Reconsideration—After NonFinal Reject issued on Aug. 28, 2017 by the USPTO for U.S. Appl. No. 14/352,107, which was filed on Apr. 16, 2014 and granted as U.S. Pat. No. 9,790,555 on Oct. 17, 2017 (Inventor—Chunde Li, et al.) (4 pages). |
| Issue Notification issued on Sep. 27, 2017 by the USPTO for U.S. Appl. No. 14/352,107, which was filed on Apr. 16, 2014 and granted as U.S. Pat. No. 9,790,555 on Oct. 17, 2017 (Inventor—Chunde Li, et al.) (1 paqe). |
| VanGuilder et al. Biotechniques 44:619-626 (Year: 2008). |
| Shariat et al., Journal of Clinical Oncology, vol. 20, No. 3, Feb. 1, 2002: pp. 833-841. |
| Strijbos et al., European Journal of Cancer 2010;46(1):2027-2035. |
| Kaushal et al., Appl. Immunohistochem. Mol. Morphol., vol. 16, No. 1, Jan. 2008, pp. 1-6. |
| Ohta et al., Anticancer Research 2002;22:2991-2996. |
| Nakagawa et al., “A Tissue Biomarker Panel Predicting Systemic Progression after PSA Recurrence Post-Definitive Prostate Cancer Therapy,” PLoS ONE 2008;3(5): e2318. doi: 10.1371/journal.pone.0002318. |
| Thangaiah et al., N A J Med. Sci. 2014;7(2):75-80. |
| Number | Date | Country | |
|---|---|---|---|
| 20210017606 A1 | Jan 2021 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 15725678 | Oct 2017 | US |
| Child | 16904789 | US | |
| Parent | 14352107 | US | |
| Child | 15725678 | US |
