Methods for analyzing allelic expression of PIK3CA in cancer and uses thereof

Abstract

The present invention provides a method for identifying a subset of cancer patients with poor prognosis, comprising comparing allelic expression of mutant and wild-type PIK3CA alleles in patient samples. The present invention further teaches the use of drugs that target mutant PIK3CA proteins in the subset of patients determined to express mutant PIK3CA alleles above a certain threshold, as well as kits or systems capable of assessing PIK3CA allele expression.

Description

[Technical field]

[0001] The present invention provides methods for determining the allelic expression of the oncogene PIK3CA as a biomarker of breast cancer outcome or to identify subsets of patients likely to respond to treatment with PI3K inhibitors.

[0002] This application benefits from the priority of Portuguese Application No. 20211000008758 and European Application No. 21161118.1, each filed on March 5, 2021, both of which are incorporated herein by reference. [Background technology]

[0003] Humans are diploid organisms, and it is possible for the two alleles of a gene locus to be expressed in different proportions in heterozygous individuals, which is called allelic imbalance. Allelic imbalance in gene expression can arise from single base changes that regulate gene expression levels, such as germline mutations (Figure 1A), cis-regulatory mutations, somatic mutations, and allele-specific epigenetic events. Normal cis-regulatory variations are known to be the major determinants of allelic expression imbalance, affecting widespread portions of the human genome and generating a wealth of phenotypic intraspecies variation. Cis-regulatory variations alter the activity of cis-regulatory elements by affecting protein binding to DNA, protein RNA, and miRNA. The most robust approach to detect their effects and map them is by analyzing differential allelic expression (DAE) in individuals heterozygous for transcriptional polymorphisms. PIK3CA, one of the most frequently mutated genes in breast cancer, has been shown to display imbalanced expression of mutant alleles in a small proportion of lung cancers, with copy number gain being the most frequent alteration associated with this increased expression. Summary of the Invention [Problem to be solved by the invention]

[0004] Based on the above-mentioned state of the art, the object of the present invention is to provide means and methods for determining the allelic expression of mutant and wildtype alleles of the PIK3CA gene in tumor samples obtained from heterozygous cancer patient subjects in order to predict prognosis and / or sensitivity to certain drugs. [Means for solving the problem]

[0005] This object is achieved by the subject matter of the independent claims herein, together with further advantageous embodiments described in the dependent claims herein, the examples, the figures and the general description. [Brief description of the drawings]

[0006] [Figure 1]Figure 1 shows cis-regulatory variations of PIK3CA gene expression in normal breast tissue. A: Schematic of cis-regulatory SNPs (rSNPs) that result in DAE on the target gene. Different alleles of rSNPs may have differences in binding affinity with transcription factors (red), resulting in differences in the expression of both alleles, the effect of which can be measured in individuals heterozygous for the daeSNP. B: DAE ratios of six daeSNPs in the PIK3CA gene region. Each dot is each of the heterozygotes for the corresponding variant shown on the x-axis, and the dotted lines delimit the 1.5-fold difference in preferential allele expression (|DAE|=0.58); C: rs2699887 is significantly associated with the DAE of rs12488074. The p-values ​​shown correspond to Student's t-test with 1000 permutations; D: Genomic map of the location of rs2699887 showing methylation of surrounding CpG sites (MeDIP-seq), typical active promoter chromatin modification marks (H3K4me3, H3K9ac, H3K36me3), and open chromatin state (DNAseI hypersensitivity and chromatin accessibility p-values ​​(p-val)) as derived from the Breast HMEC Epigenomics Project Roadmap. [Figure 2-1] Figure 2 shows that mutant allele imbalance in gene expression of somatic missense PIK3CA mutations is frequent in breast tumors, with the mutant allele being preferentially expressed. A: Schematic of the hypothesis that cis-acting regulatory variants (rVars), either germline derived or somatically acquired, result in different relative allelic expression of mutant and wild-type alleles, resulting in tumors with different prognosis. B: Distribution of α, β, γ ratios in breast tumors. Tumors with mutant allele differential expression (MADE), |γ| ≥ 0.58, are shown in red. [Figure 2-2]C: Correlation analysis of alpha vs. beta and alpha vs. gamma, showing that both genomic copy number dosage and allelic expression regulation contribute to allelic imbalance in the expression of tumor mutations. D: Comparison of matched beta and gamma values, showing the predominance of tumors with high copy number of the wild-type allele but preferential allelic expression of the mutant allele. [Figure 3-1] Figure 3 shows the survival and clinicopathological effects of preferential expression of the mutant allele. A: Kaplan-Meier curves of disease-specific survival in METABRIC, showing a worse prognosis for patients with differential expression of PIK3CA mutations (MADE group, shown in yellow) compared to patients expressing equimolar levels of mutant and wild-type alleles (mut=wt group, shown in blue). The bottom of the graph shows the number of patients at risk per group over time. B: Kaplan-Meier curves of disease-specific survival for the MADE group, subdivided into groups with preferential expression of the mutant allele (MADE_mut, shown in red) and groups with preferential expression of the wild-type allele (MADE_wt, shown in green) in METABRIC, confirming the worse survival compared to the mut=wt group (shown in blue). The bottom of the graph shows the number of patients at risk per group over time. [Figure 3-2] C: Preferential expression of the mutant allele is associated with ER-negative, PR-negative, and Her2-positive breast tumors. In all graphs, samples are colored according to the MADE classification. P values ​​correspond to Wilcoxon rank sum test with continuity correction and corrected for multiple testing using the Bonferroni correction. [Figure 3-3] D: Kaplan-Meier curves showing worse disease-specific survival in the MADE group compared with the mut=wt group in ER-positive, PR-positive, and Her2-negative subtypes in METABRIC (ER = estrogen, PR = progesterone, Her2 = human epidermal growth factor 2 receptor). [Diagram 3-4]D: Kaplan-Meier curves showing worse disease-specific survival in the MADE group compared with the mut=wt group in ER-positive, PR-positive, and Her2-negative subtypes in METABRIC (ER = estrogen, PR = progesterone, Her2 = human epidermal growth factor 2 receptor). [Figure 4] Figure 4 shows that rs2699887 is an eQTL for PIK3CA expression in METABRIC tumors. The indicated P-values ​​correspond to Student's t-test. [Figure 5-1] Figure 5: Association analysis of alpha ratios with ER, PR, and HER2 status. Box plots of alpha ratios for METABRIC tumors (top) and TCGA tumors (bottom). Each dot represents a sample, and the number of samples in each group is given in brackets. In all graphs, samples are colored according to the MADE classification. P values ​​correspond to Wilcoxon rank sum test with continuity correction and corrected for multiple testing using the Bonferroni correction. [Figure 5-2] Figure 5: Association analysis of alpha ratios with ER, PR, and HER2 status. Box plots of alpha ratios for METABRIC tumors (top) and TCGA tumors (bottom). Each dot represents a sample, and the number of samples in each group is given in brackets. In all graphs, samples are colored according to the MADE classification. P values ​​correspond to Wilcoxon rank sum test with continuity correction and corrected for multiple testing using the Bonferroni correction. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0007] Summary of the Invention The examples presented herein demonstrate the clinical impact of differential allele expression in tumors with somatic mutations of PIK3CA. DNA and RNA sequencing data from two independent sets of breast tumors, from a total of 255 breast cancer cases, show frequent preferential expression of mutant alleles compared to healthy controls. Differential expression of mutant alleles can be associated with different clinicopathological variables in breast tumors or define subsets of patients with different survival rates.

[0008] A first aspect of the present invention provides a method for prognostic stratification of patients diagnosed with a PIK3CA mutated cancer, the method comprising determining the expression levels of both wild-type and mutant alleles of the PIK3CA gene present in a tumour sample and assigning various clinical outcomes based on the relationship between these two values.

[0009] In another embodiment of the present invention, the above defined method is used to identify breast cancer patients who are likely to respond favorably to treatment with phosphoinositide-3-kinase (PIK3) inhibitors due to unbalanced expression of mutant PIK3CA alleles.

[0010] The present invention further teaches the use of kits or systems capable of assessing PIK3CA allele expression, as well as drugs that target mutant PIK3CA proteins in subsets of patients in which expression of mutant PIK3CA alleles above a certain threshold is determined.

[0011] Detailed Description of the Invention Terms and Definitions For the purposes of interpreting this specification, the following definitions shall apply, and where appropriate, terms used in the singular shall include the plural and vice versa. In the event that a definition set forth below conflicts with any document incorporated herein by reference, the definition set forth herein shall control.

[0012] As used herein, the terms "comprising," "having," "containing," "including," and other similar forms and their grammatical equivalents are intended to be equivalent in meaning and to be open-ended in that the item or items following any one of these words are not intended to be an exhaustive list of such item or items or are not limited to only the listed item or items. For example, an item "comprising" components A, B, and C can consist of components A, B, and C (i.e., contain only components A, B, and C), or can include not only components A, B, and C, but also one or more other components. Thus, "comprising" and similar forms and their grammatical equivalents are intended and understood to include disclosure of embodiments that "consist essentially of" or "consist of."

[0013] Where a range of values ​​is provided, unless otherwise defined, it is understood that each intervening value, to the tenth of the unit of the lower limit, between the upper and lower limit of that range, and any other stated or intervening value within that stated range, is encompassed within the disclosure, subject to any specifically excluded limit of the stated range. Where one or both of the limits are included in the stated range, ranges excluding either or both of those included limits are also included in the disclosure.

[0014] As used herein, reference to "about" a value or parameter includes (and describes) the variance directed to the value or parameter itself. For example, a statement referring to "about X" also includes the statement "X."

[0015] As used in this specification, including the appended claims, the singular forms "a," "or," and "the" include plural references unless the context clearly dictates otherwise.

[0016] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art (e.g., cell culture, molecular genetics, nucleic acid chemistry, hybridization techniques, and biochemistry). Standard techniques are used for molecular biology, genetic and biochemical methods (see generally Sambrook et al., Molecular Cloning: A Laboratory Manual, 4th ed. (2012) Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY and Ausubel et al., Short Protocols in Molecular Biology (2002) 5th ed., John Wiley & Sons, Inc.) and chemical methods.

[0017] The term "gene" refers to a polynucleotide that contains at least one open reading frame (ORF) that can encode a particular polypeptide or protein after being transcribed and translated. A polynucleotide sequence can be used to identify larger fragments or full-length coding sequences of the gene with which it is associated. Methods for isolating larger fragment sequences are known to those of skill in the art.

[0018] The term "genotype" in this specification refers to the number of copies of either wild-type or mutant alleles present at a particular gene or locus.In other words, a sample or patient is heterozygous or homozygous for a particular nucleic acid sequence, for example, mutant PIK3CA gene.Homozygous cancer patient samples have two identical alleles, for example, two wild-type non-mutated PIK3CA alleles, while heterozygous samples have multiple types of alleles, including wild-type and mutant alleles.

[0019] The term "gene expression" or "expression", or the term "gene product", can refer to either or both of the process and product of the production of a nucleic acid (RNA) or a peptide or polypeptide, also called transcription and translation, respectively, or any intermediate process that controls the processing of genetic information to produce a polypeptide product. The term "gene expression" can also apply to the transcription and processing of RNA gene products, such as regulatory RNA or structural (e.g., ribosomal) RNA. If the expressed polynucleotide is derived from genomic DNA, expression can include splicing of mRNA in eukaryotic cells. Expression can be measured at both the transcriptional and translational levels, in other words, the mRNA and / or the protein product.

[0020] The term "phosphoinositide-3-kinase, catalytic, alpha polypeptide (PIK3CA)" in the context of this specification relates to the PIK3CA gene (ENSG00000121879) encoding the phosphatidylinositol kinase 3 protein p110α subunit (UniProt number P42336). Mutations within external regulatory elements, regulatory intron regions, or the 5' and 3' untranslated regions of this gene have been shown to affect allelic expression of PIK3CA mRNA, and regions within 250 Kb upstream or downstream of the PIK3CA open reading frame are considered to be encompassed by the term "PIK3CA" or "PIK3CA gene" for the purposes of the present invention.

[0021] The term "wild type" or "wild type PIK3CA allele" according to the present invention refers to the nucleic acid sequence of the PIK3CA gene that is free of mutations (ENSG00000121879).

[0022] The terms "PIK3CA mutation", "PIK3CA variant" or "variant" in the context of this specification refer to a sequence change or modification within the transcribed PIK3CA locus, with the Ensembl database entry for the PIK3CA gene currently listing 10 known splice variants and 344 known mutant PIK3CA allele nucleic acid sequences. PIK3CA mutations of particular relevance to the present invention are those present in PIK3CA mRNA transcripts associated with cancer phenotypes, which are generally missense mutations that result in altered transcription, translation, or activity of the PIK3CA oncogene. Most of the known PIK3CA mutations are single nucleotide polymorphisms (SNPs).

[0023] The term "single nucleotide polymorphism (SNP)" in the context of this specification relates to a single base that differs from the wild-type sequence, e.g., from the wild-type PIK3CA gene as defined above. This term encompasses variants or mutations detected in a single nucleic acid within the open reading frame of PIK3CA DNA, or within 250 kb of this gene, or within the mRNA transcribed from said nucleic acid location. SNPs of particular interest according to the invention are identified in the method of differential gene expression analysis provided in the Examples. SNPs that are particularly useful as biomarkers according to the invention are regulatory SNPs such as rs2699887 that are associated with allelic imbalance in PIK3CA, the six differentially expressed SNPs identified in Table 1, and those in close linkage disequilibrium with the SNPs of Table 1, such as those listed in Tables 3 and 4, all of which contribute to differential allelic expression at the PIK3CA locus that is utilized in the predictive power of patient outcome according to the invention.

[0024] In the context of this specification, the term "allele" refers to two copies of a locus present in the genome. Humans, being diploid organisms, have two alleles at most positions, one inherited from each parent, although somatic mutations in alleles may occur later in life. Heterozygous PIK3CA alleles, in which one copy of a locus contains a variation or mutation that differs from the wild-type sequence, are of particular importance for the method according to the invention. Mutant or variant alleles include mutations, such as SNPs, at a particular locus in the genome. For example, one example of a mutant PIK3CA allele relevant to the present invention is a copy of the PIK3CA gene that has a SNP associated with poor prognosis in breast cancer.

[0025] "Mutant PIK3CA alleles" encompass variants or mutations in cis-regulatory elements within introns or untranslated regions of PIK3CA mRNA transcripts that are associated with pathogenic expression and therefore can be used as supplementary or sole biomarkers for severe forms of breast cancer.

[0026] The term "allele-specific expression" or "allelic expression" is a measure of gene expression due to one copy, or allele, of PIK3CA at a particular locus. "Allelic expression" depends on the genotype, i.e., whether a subject is homozygous or heterozygous at a particular nucleic acid region or single base pair position, as well as the copy number, i.e., the relative copy number of each allele, and other forms of regulation that may affect the transcription rate of one or both copies of the gene. Allelic expression can be defined as the ratio or relative expression comparing matching alleles, which in most cases are mutant and wild-type alleles at the same locus in heterozygous individuals.

[0027] In the context of this specification, "allelic imbalance" or "differential allele expression" refers to the two alleles of a locus in a heterozygous individual that are expressed in different ratios. A locus can be said to exhibit "allelic imbalance" if one allele is expressed more than would be expected based on copy number, and this imbalance can be calculated by normalizing the expression level of each allele to the amount expected based on genotype, or copy number in the genome. For example, considering the total amount of PIK3CA mRNA transcripts in a cell or tumor sample obtained from a heterozygous subject, in the absence of allelic imbalance, half of the total amount of PIK3CA mRNA present would come from mutant PIK3CA alleles and half would come from wild-type PIK3CA alleles. In other words, the allelic expression levels determined for each copy of the PIK3CA gene should be approximately equal since there is one copy of each in the cell. However, if significantly more than half of the total PIK3CA mRNA transcripts of a locus come from one allele, this is defined as PIK3CA allelic imbalance.

[0028] "Mutant allele imbalance" or "mutant allele differential expression (MADE)" in the context of this specification refers to mutant alleles, in particular mutant PIK3CA alleles, that are expressed at higher levels than their wild-type counterparts. In other words, it is a form of allelic imbalance as defined above, in which mutant allele transcripts, or mutant allele proteins, are more numerous than their wild-type counterparts.

[0029] As used herein, the term "pharmaceutical composition" refers to a compound of the present invention or a pharma- ceutically acceptable salt thereof, together with at least one pharma- ceutically acceptable carrier. In certain embodiments, the pharmaceutical composition according to the present invention is provided in a form suitable for topical, parenteral, or injectable administration.

[0030] As used herein, the term "pharmaceutically acceptable carrier" includes any solvents, dispersion media, coatings, surfactants, antioxidants, preservatives (e.g., antibacterial, antifungal), isotonicity agents, absorption delaying agents, salts, preservatives, drugs, drug stabilizers, binders, excipients, disintegrants, lubricants, sweeteners, flavoring agents, dyes, and the like, and combinations thereof, as known to those of skill in the art (see, e.g., Remington: the Science and Practice of Pharmacy, ISBN 0857110624).

[0031] As used herein, the term "treating" or "treatment" of any disease or disorder (e.g., cancer) refers, in one embodiment, to ameliorating the disease or disorder (e.g., delaying, preventing, or alleviating the onset of the disease or at least one of its clinical symptoms). In another embodiment, "treating" or "treatment" refers to alleviating or improving at least one physical parameter, including those that may not be discernible to the patient. In yet another embodiment, "treating" or "treatment" refers to modulating the disease or disorder, either physically (e.g., stabilization of a discernible symptom), physiologically (e.g., stabilization of a physical parameter), or both. Methods for assessing the treatment and / or prevention of a disease are generally known in the art, unless otherwise described herein below.

[0032] Despite the high frequency of PIK3CA mutations in breast cancer, responses to PI3K inhibitor treatment have not been as successful as expected, and the prognostic significance of PIK3CA somatic mutation expression in breast tumors is unclear. Although PIK3CA mutations are known to be well associated with clinical outcome and treatment response, especially in HR+ breast cancer, we found that not all PIK3CA somatic mutations are equally expressed in tumors, and some express few mutations identified at the DNA level. The findings herein highlight the importance of assessing the expression levels of mutant PIK3CA alleles as part of clinical management.

[0033] A first aspect of the present invention is a method for predicting the prognosis of a cancer patient by determining the allelic expression of wild-type and mutant PIK3CA alleles in a tumor disease characterized by a PIK3CA mutation.

[0034] The term "tumor" disease according to this aspect of the present invention encompasses neoplastic cells characterized by at least one PIK3CA mutation. Tissue-derived cancers are particularly relevant for analysis by this method, since PIK3CA mutations are frequently observed across a wide range of different tumor types, such as epithelial cell-derived cancers, such as melanoma, lung cancer, and endometrial cancer (Morgese F. 2014, Oncotarget, 8:75914; Li Q. 2013, Clin. Cancer Res, 19:6252; Wang L., 2012, J. Cancer Res. Clin. Oncol. 138:377).

[0035] Particular embodiments of this aspect of the invention relate to tumour samples, such as biopsies, or tissue samples containing malignant cells, for example lymph nodes draining a tumour site.

[0036] Certain embodiments relate to samples obtained from patients suspected of having or diagnosed with breast cancer, such as biopsies, or breast cancer metastatic sites, such as resected lymph node metastases. The expression status of PIK3CA according to these embodiments provides information regarding the allelic expression of both inherited and somatic mutations present in the cancer cells.

[0037] The method according to this aspect of the invention comprises carrying out the following steps on a tumour sample containing genetic material obtained from said patient:

[0038] In a first measuring step, the following values ​​are determined for the sample: i. the expression (ex) level of the mutant (mut) PIK3CA allele (ex(mut)), and ii. Expression (ex) levels of wild-type (wt) PIK3CA alleles (ex(wt)) Here, a mutant PIK3CA allele may be any mutation in the PIK3CA gene, and in particular may be a SNP located in the PIK3CA gene, such as those listed in Tables 1, 3, and 4 of the Examples.

[0039] In the allocation step, the expression levels of wild-type and mutant PIK3CA alleles are compared to provide evidence of imbalance, or differential allele expression, if either the mutant or wild-type allele is significantly overexpressed relative to the other. The ratio of wild-type to mutant PIK3CA allele expression is calculated (allelic expression value), expressed as a fold change value, and this value is used to assign a clinical outcome to the patient. An allelic expression, or fold change, of 1 indicates equivalent levels of wild-type and mutant allele expression. The threshold for significant overexpression of one allele relative to another, indicating allelic imbalance, should be precisely discriminative between patient groups with different PIK3CA allele expression that correlate with distinct survival outcomes, based on statistical methods such as those shown in Figure 3 of the Examples. The range, or threshold, that defines a significant difference in the expression levels of the two alleles being compared should be determined keeping in mind the error rate of the measurement method used and the range of ratios present in a cohort of patient samples obtained from the same disease as the known outcome.

[0040] According to this aspect of the invention, if the expression level of either the wild-type PIK3CA allele or the mutant PIK3CA allele is significantly higher than the expression level of the other, particularly where the wild-type PIK3CA allele is significantly more expressed than the mutant PIK3CA allele, in other words, where an allelic imbalance is observed at the PIK3CA locus, the patient is assigned a poor prognosis.

[0041] Certain embodiments relate to assigning a prognosis to patients with equivalent expression of PIK3CA mutant and wild-type alleles, where poor prognosis is defined as about an 11-year survival probability of about 45-55% or less and good prognosis is defined as about an 18-year survival probability of about 75-85%.

[0042] Another aspect of the invention relates to a method comprising measuring the expression levels of wild-type and mutant PIK3CA alleles as defined in the first aspect of the invention, and using a threshold of a 1.5-fold difference in expression of mutant or wild-type PIK3CA alleles, assigning a patient as having significant allelic overexpression or imbalance for the purposes of assigning a prognosis according to the invention.

[0043] Patients whose tumor samples are determined to be characterized by allelic expression imbalance, where either mutant or wild-type PIK3CA is expressed at least 1.5-fold more than the other, are assigned a poor prognosis according to this aspect of the invention, particularly where the mutant allele is expressed at least 1.5-fold more than the wild-type. Patients whose tumors are characterized by the absence of PIK3CA allelic imbalance according to this aspect of the invention, where each PIK3CA allele is expressed equally, where the wild-type and mutant allele expression levels differ from each other by less than 1.5-fold, are assigned a more favorable breast cancer clinical outcome compared to the poor prognosis allelic imbalance group defined above.

[0044] Some embodiments of this aspect of the invention that are particularly relevant to breast cancer relate to assigning patients who express levels of wild-type and mutant PIK3CA alleles that differ from each other in tumor samples less than 1.5-fold to a favorable prognosis, where favorable prognosis is defined as about a 75-85% probability of survival over about 18 years.

[0045] Some embodiments of this aspect of the invention, particularly relevant to breast cancer, relate to assigning a poor prognosis, defined as about a 50-60% chance of survival about 11 years, to patients whose tumor samples are determined to have expression of either mutant or wild-type PIK3CA alleles, and in particular expression of the mutant allele 1.5-fold or greater.

[0046] In some embodiments of this aspect of the invention, patients whose tumor samples are determined to exhibit overexpression of the wild-type allele, and in particular, those determined to exhibit a threshold of 1.5 times or greater levels of the PIK3CA mutant allele, are assigned a very poor prognosis, as this subset of patients has been shown to have the poorest overall outcomes in breast cancer.

[0047] Some embodiments of this aspect of the invention that are particularly relevant to breast cancer relate to assigning a poor prognosis to patients having tumor samples assessed as overexpressing the wild-type PIK3CA allele relative to the mutant allele, where poor prognosis is defined as having about a 45-55% probability of survival at about 5.5 years.

[0048] Figure 5 in the Examples shows that overexpression of mutant PIK3CA alleles is associated with poor prognosis compared to patients in which each allele is expressed at equivalent levels, and overexpression of wild-type PIK3CA alleles is associated with even poorer patient survival based on a 1.5-fold threshold as a definition of allelic imbalance.

[0049] One skilled in the art will understand that an appropriate threshold of overexpression can be identified by analyzing a patient cohort according to the methods used in the Examples to assess allelic imbalance in breast cancer tumor samples, and that the threshold may vary slightly depending on the mutation being measured and the methodology selected for measuring PIK3CA allele expression levels according to the present invention.

[0050] In some embodiments, the threshold for significant allelic overexpression or allelic imbalance according to the invention is a range encompassing statistically equivalent expression of the two alleles, such as the range calculated from a cohort of comparable breast cancer samples in the Examples. In another embodiment, the sample is compared to a control reference sample that is pre-determined to represent the limits of this range.

[0051] In one embodiment of the above aspect of the invention, the mutant PIK3CA allele is determined at the level of the full mRNA transcript containing the particular PIK3CA mutation.

[0052] In another embodiment of the method according to the above aspect of the invention, the expression level of PIK3CA mutations is assessed at the level of single base changes, in other words SNPs.

[0053] Each SNP in Table 1 was evaluated for its association with breast cancer patient outcomes, and each was shown to represent allelic imbalance in certain patients. It is possible to combine allelic expression information from multiple SNPs in cancer samples that are heterozygous for multiple PIK3CA mutations or when evaluating the expression levels of SNPs located in both exonic and intronic regions.

[0054] In some embodiments of the methods according to the aspects of the invention defined above, the assay for determining PIK3CA allele expression levels is performed on a tumor sample not previously determined to contain a mutant PIK3CA allele as part of the step of screening for the presence of a mutant PIK3CA allele. This can be done by performing an assay that simultaneously assesses the expression levels of multiple mutant PIK3CA alleles, for example using a microarray to assess the expression levels of transcripts containing different SNPs, as shown in the examples.

[0055] Another embodiment of the method according to the above aspect of the invention relates to the use of any quantitative allele-specific measurement of mRNA expression levels, such as mRNA sequencing of the PIK3CA gene and quantification of transcripts having either wild-type or mutant PIK3CA sequences, or by the use of primers suitable for polymerase chain reaction amplification of specific loci characterized by mutant or wild-type alleles.

[0056] Another aspect of the present invention provides a method for determining the likelihood of response of a tumor disease in a cancer patient, particularly a breast cancer patient, to treatment with phosphoinositide-3-kinase (PIK3) inhibitors, which identifies patients who overexpress mutant PIK3CA alleles and are therefore more likely to express mutant proteins targeted by these drugs.

[0057] As used herein, "PIK3 inhibitors" or "PIK3 inhibitor therapy" refers to pharmaceutical compounds that inhibit class I phosphoinositide-3-kinases, including the α subunit encoded by the PIK3CA gene. PIK3 inhibitors that are particularly useful for the purposes of the present invention include, but are not limited to, compounds such as alpelisib (Piqray), which targets α subunit mutations, or copanlisib (Aliqopa), which targets α and β subunit mutations.

[0058] These drugs target the mutant variants of PI3K protein and therefore will only work if the mutant protein is transcribed, but the inventors have found that this is not the case for all breast cancer patients. The inventors believe that tumors that overexpress mutant PIK3CA alleles are more likely to respond to the chemotherapy compared to heterozygous patients who express wild-type and mutant alleles equally, and even more likely to respond compared to patients who express less mutant alleles.

[0059] In a first measuring step, the allelic expression of mutant and wild-type PIK3CA alleles is assessed according to the first aspect of the invention, in other words, expression levels for mutant and wild-type PIK3CA gene alleles in a patient sample are obtained using a quantitative allele-specific methodology to compare the relative abundance of mutant and wild-type transcripts.

[0060] In a second assignment step, the subject is assigned a likelihood of response of the cancer patient's tumor disease, particularly the breast cancer patient's tumor disease, to treatment with a mutant PI3K inhibitor drug if the mutant PIK3CA allele is significantly overexpressed compared to the wild-type allele at that locus, particularly where the mutant PIK3CA allele expression value is 1.5-fold greater than the wild-type allele.

[0061] Establishing statistical significance thresholds between PIK3CA expression status values ​​is demonstrated in examples examining patient outcomes and allelic imbalance in cohorts of similar patients. It is understood that the exact value of the threshold may be subject to change as more samples are analyzed or if methods other than mRNA microarrays are used to measure allelic imbalance in accordance with the present invention.

[0062] The method identifies patients with mutant PIK3CA alleles expressed at the mRNA level and uses the value to stratify patients into groups that may have different outcomes from PIK3 inhibitory drugs as defined above. Here, patients assigned a higher likelihood of response include both patients whose PIK3CA expression status values ​​are within a range indicating equal expression of wild-type and mutant alleles, and also patients who overexpress the mutant allele. These two groups, especially the latter, are predicted to respond better to drugs targeting the mutant than those with low mRNA expression of mutant PIK3CA, which results in low protein expression of the mutant allele. This information can help clinicians identify patients who are more likely to benefit from PI3K-targeted chemotherapy (prolonged overall or disease-specific survival).

[0063] The manner in which the expression level of PIK3CA alleles is measured at the mRNA level is not limited by the present invention and includes using allele-specific nucleic acid probes, particularly using quantitative PCR methods such as real-time PCR or qPCR, sequencing reactions, or nucleic acid arrays. In other embodiments, the expression status of PIK3CA is determined by mRNA sequencing.

[0064] One particularly useful methodology for measuring the expression level of biomarker PIK3CA nucleic acid alleles is a nucleic acid amplification method using polymerase chain reaction of RNA extracted from patient tumor samples. The term "cycle threshold" in the context of this specification relates to quantitative nucleic acid measurement, for example, measurement by "quantitative polymerase chain reaction (qPCR)". This method involves repeated cycles of nucleic acid amplification using nucleic acid probes that hybridize with the target wild type and mutant alleles to generate products that emit fluorescent signals, and measuring the products to determine the amount of starting genetic material. The cycle threshold may be an average value or an average value of a number of replicate samples. Other quantitative measurements, such as crossing points or adjusted inflection points, may be substituted for the cycle threshold.

[0065] Those skilled in the art will understand that in embodiments where PIK3CA allele variant expression is measured by qPCR, the values ​​are the expression cases of cycle threshold difference compared to housekeeping genes, i.e., the number of qPCR cycles required to generate a fluorescent signal from the specific nucleic acid probe used that exceeds a user-defined threshold. Thus, these values ​​reflect PCR conditions and cycle thresholds, and the exact pathogenic expression threshold value can be expected to differ from the case obtained from mRNA sequencing samples. Similar thresholds related to breast cancer outcomes for qPCR measurement of allele expression can be obtained by evaluating a cohort of similar samples using the same methodology as performed in the case and performing correlation analysis with patient outcomes. Specific nucleic acid probes can identify the expression of PIK3CA variants, for example, using standard TAQman ABI or Sybrgreen enzyme qPCR assay conditions, using primer designs that include sequences complementary to either WT (wild type), or SNP regions, or alleles.

[0066] In another embodiment of the method for predicting the prognosis or drug response of cancer patients according to the above-defined aspects, the allelic expression levels of mutant and wild-type PIK3CA copies are evaluated by mRNA microarray assays, such as the Illumina array used in the examples, or by mRNA sequencing technology. These methodologies allow for non-targeted evaluation of any mutant alleles present in the patient sample, since the patient may be heterozygous for multiple SNPs.

[0067] In another embodiment of the method for establishing the prognosis or drug sensitivity of a cancer patient according to the above defined aspects of the invention, the expression levels of mutant PIK3CA alleles are determined for at least one SNP locus selected from rs7636454, rs3960984, rs3729679, rs2699887, rs12488074, rs4855093 and / or rs9838411. In embodiments where the mutation status of a patient sample is not yet known, it is preferred that the expression levels of mutant PIK3CA alleles are determined for several common SNP loci, or for all the listed loci, since some SNPs are relatively rare and will be detected only infrequently in patients.

[0068] In certain embodiments of the above-mentioned aspects of the invention, the expression levels of mutant PIK3CA alleles and wild-type PIK3CA alleles are determined for common SNP mutations found in more than half of patients with mutant PIK3CA gene, such as SNP locus rs3960984. Measuring the expression levels of the alleles of SNP rs3960984 and its wild-type counterpart is particularly useful according to the invention, both due to the high occurrence of unbalanced allele expression of mutant and wild-type rs3960984 alleles, as well as its high prevalence in PIK3CA mutant breast cancer samples (Table 1). Similar common SNPs among other known cancer patients are particularly desirable for use according to the invention.

[0069] In another embodiment of the above defined aspect of the invention, the expression levels of mutant and wild type alleles are determined at a SNP locus located within an untranslated region of the PIK3CA transcript, in particular where the expression level of mutant PIK3CA alleles is determined for SNP rs2699887. The allelic imbalance of some mutant PIK3CA alleles rs12488074 is shown in Figure 1 of the Examples to be influenced by the presence of the cis-regulatory SNP locus rs2699887.

[0070] Certain embodiments of the above defined methods relate to predicting prognosis or tumor sensitivity to a PIK3 inhibitor in a patient diagnosed with a form of breast cancer. In some embodiments of the invention according to any of the previous aspects of the invention, the patient has been diagnosed with a breast cancer tumor classified as estrogen receptor positive, progesterone receptor positive, or human epidermal growth factor receptor 2 negative.

[0071] In the data presented in the Examples, preferential expression of mutant PIK3CA alleles is associated with known poor prognostic variables such as PR-negative and Her2-positive status. In the METABRIC dataset, the inventors found that patients with tumors with mutant allele expression imbalance have poorer overall survival than those without, especially in the ER-positive, PR-positive, and Her2-negative subsets. The method according to the invention can provide additional information in comparison with the current standard of care prognostic histological markers by identifying subsets of patients with different survival outcomes (Figure 3).

[0072] Another aspect of the present invention relates to a system enabling the method as described above, comprising a device for obtaining expression data from tumor cells, a sequencing apparatus for generating PIK3CA allele expression data, and a device for electronically processing such PIK3CA allele expression data according to the present invention as disclosed herein.

[0073] Pharmaceutical Compositions and Administration Another aspect of the present invention relates to a pharmaceutical composition comprising a PIK3 inhibitor selected from alpelisib, or copanlisib, or a pharma- ceutically acceptable salt thereof, and a pharma- ceutically acceptable carrier. In a further embodiment, the composition comprises at least two pharma- ceutically acceptable carriers, such as those described herein.

[0074] In certain embodiments, alpelisib or copanlisib is administered in combination with an additional cancer medicament in a combination therapy, for example in combination with a selective estrogen receptor degrader, such as fulvestrant.

[0075] In certain embodiments of the present invention, the compounds of the present invention are typically formulated into pharmaceutical dosage forms to provide easily controllable administration of the drug and to provide the patient with a well-defined, easy to handle product.

[0076] The pharmaceutical compositions can be formulated for enteral administration, particularly for oral or rectal administration. Furthermore, the pharmaceutical compositions of the present invention can be in a solid form (including, but not limited to, capsules, tablets, pills, granules, powders, or suppositories) or in a liquid form (including, but not limited to, solutions, suspensions, or emulsions).

[0077] The pharmaceutical composition may be formulated for parenteral administration, for example by intravenous (iv) injection, intradermal, subcutaneous or intramuscular administration.

[0078] The dosing regimen of the PIK3 inhibitor compounds of the present invention will vary according to known factors such as the pharmacodynamic properties of the particular agent and its mode and route of administration; the recipient's species, age, sex, health, medical condition, and weight; the nature and extent of symptoms; type of concurrent treatment; frequency of treatment; route of administration, renal and hepatic function of the patient, and the desired effect. In certain embodiments, the compounds of the present invention may be administered once daily, or the total daily dosage may be administered in divided doses two, three, or four times daily.

[0079] In certain embodiments, the pharmaceutical composition containing the PIK3 inhibitor of the present invention may be a unit dose of about 1-1000 mg of active ingredient(s) for a subject weighing about 50-70 kg. The therapeutically effective dose of the compound, pharmaceutical composition, or combination thereof depends on the subject's species, weight, age, individual condition, and the disorder or disease being treated or its severity. A physician, clinician, or veterinarian of ordinary skill can readily determine the effective amount of each active ingredient required to prevent, treat, or inhibit the progression of a disorder or disease.

[0080] One aspect of the invention relates to a PI3K inhibitor for use in treating a breast patient determined to be responsive by a method according to any of the previously described aspects of the invention, particularly to a PI3K inhibitor selected from alpelisib or copanlisib.

[0081] A further aspect of the invention is a kit for analyzing biomarkers for determining PIK3CA expression status in a patient tumor sample comprising: a. a mutant PIK3CA allele, and b. Wild-type PIK3CA allele; The kit includes the use of a nucleic acid probe to amplify and / or detect the expression level of a PI3K inhibitor. The kit can be used in tumor samples from patients to provide clinicians with information regarding the prognosis of breast cancer patients or to predict the likelihood that a breast cancer patient will respond well to treatment with a PI3K inhibitor.

[0082] The method may be performed by a computer-implemented method, in particular where the steps of evaluating and allocating are performed by a computer.

[0083] Furthermore, the method can be implemented by a computer program comprising computer program code that, when executed on a computer, causes the computer to perform at least the evaluating and / or assigning steps. In particular, the results of the measuring step can be provided to the computer and / or the computer program by user input and / or by providing a computer readable file that includes information regarding the methylation levels obtained during the measuring step. The results from the measuring step can be stored in the memory of the computer, in a non-transitory storage medium for further processing.

[0084] Medical Treatments, Dosage Forms and Salts Similarly within the scope of the present invention is a method of treating breast cancer in a patient in need thereof comprising administering to the patient a PI3K inhibitor drug as described above, in particular a PI3K inhibitor drug selected from alpelisib (CAS No. 1217486-61-7) or copanlisib (CAS No. 1032568-63-0).

[0085] Those of skill in the art will recognize that a specifically mentioned drug may exist as a pharma- ceutically acceptable salt of the drug, which contains an ionized drug and an oppositely charged counterion.

[0086] Where alternative forms of a single separable feature, such as, for example, a PI3K inhibitor or a medical indication, are described herein as "embodiments," it is to be understood that such alternative forms can be freely combined to form separate embodiments of the invention disclosed herein.

[0087] The present invention is further illustrated by the following examples and figures, from which further embodiments and advantages can be derived, which are intended to illustrate the invention without limiting its scope.

[0088] Table 1 shows the DAE (differential allelic expression) of normal breast samples at six daeSNPs located in the PIK3CA region. Table 2 summarizes the allelic expression imbalance of PIK3CA somatic mutations in breast tumors. Table 3 shows the linkage disequilibrium between daeSNPs of PIK3CA. Data were mapped and annotated to European populations using the SNiPA (https: / / snipa.helmholtz-muenchen.de) tool according to the “grch37 / 1kgpp3v5” release. Table 4 lists the candidate SNPs obtained from mapping analysis of the DAE data in PIK3CA. Table 5 shows summary statistics of MADE survival analysis of PIK3CA in the TCGA set. Median, LCL and UCL are in days, LCL (lower confidence level): lower confidence level, UCL (upper confidence level): upper confidence level, N: number of people in each group. EXAMPLES

[0089] Materials and Methods Differential allelic expression analysis DNA and total RNA from 64 normal breast tissue samples were analyzed using Illumina Exon510S-Duo arrays (humanexon510s-duo) as described (Liu R., Bioinformatics 2012,28:1102). After normalization, SNPs with mean log2RNA intensity values ​​lower than 9.5 and heterozygous values ​​lower than 5 were excluded from the analysis. Two-sample Student's t-tests were applied to compare RNA log ratios between heterozygous (AB) and homozygous (AA and BB) groups. Only SNPs with p-values ​​lower than 0.05 for all comparisons were further analyzed. The following formula was used to normalize DAE: log2((RNA allele A / RNA allele B) / (DNA allele A / DNA allele B)). The analysis was performed using R with Bioconductor packages as described (Huber W., Nature Pub.Group 2015,12:115). DAE was estimated when |AE ratio| ≥ 0.58 (≥ 1.5-fold difference). This threshold was established based on the sensitivity and specificity of the applied DAE detection method. Linkage disequilibrium (LD) between daeSNPs was assessed using the genetic variant-centered annotation browser SNiPA (Li Y., Genet Epidemiol. 2010, 34:816).

[0090] Genotype imputation analysis of normal breast tissue samples Genotype imputation was performed on Illumina Exon 510 Duo germline genotype data from 64 samples that passed microarray quality control filters. Prior to imputation data, quality control was performed on the genotyping data, which required a call rate of <85%, minor allele frequency of <0.01, and p-value <1.0e. -05SNPs with Hardy-Weinberg equilibrium of were excluded from the analysis. Genotype data for chromosome 3 Illumina SNPs that passed quality control were used to impute genotypes at all known additional SNPs in the chromosome using MACH1.0 (Li, Y. et al., Genet. Epidemiol. 34, 816-834, 2010) and the staged haplotypes from the HapMap3 release (HapMap3 NCBI Build 36, CEU panel-Utah residents with Northern and Western European ancestry) (copyright) (as the reference panel). For imputation using MaCH1.0, a two-stage imputation process was used: model parameters (crossover rate and error rate) were estimated prior to imputation using all haplotypes studied, and 100 iterations of the Hidden Markov Model (HMM) were run using the command options: -greedy and -r 100. Genotypes were then imputed using the model parameter estimates from the previous round with the command options -greedy, -mle, and -mldetails. The imputation results were evaluated by a platform-specific imputation uncertainty index (rsq score) for each SNP. The MaCH-rsq score is equal to the ratio of the empirically observed variance of allele abundance to the expected binomial variance p(1-P) at Hardy-Weinberg equilibrium, where p is the observed allele frequency derived from HapMap or estimated from our own data. This value tends to zero when the uncertainty of the imputation results increases. The imputation results were evaluated by a platform-specific imputation uncertainty index (rq score) for each SNP, and results were filtered if the rq score was higher than 0.3 (Li Y., Genet Epidemiol.34:816-834,2010).

[0091] Differential allelic expression (DAE) mapping analysis of normal breast tissue samples Differential allele expression mapping analysis was performed by stratifying the AE ratios at each PIK3CA daeSNP according to genotypes at genotyped / imputed SNPs located within ±250Kb of the daeSNP. Two-sample t-tests were applied to assess the difference between the mean AE ratios between the heterozygous group samples and the combined homozygous group. P-values ​​were corrected for multiple testing using a permutation procedure of N=1000. Permutation-corrected p-values ​​<0.05 were considered significant, and on average, heterozygous samples showed a greater fold difference between alleles when compared to homozygous samples.

[0092] Proxy SNP queries Proxy SNPs were identified using HaploReg v4.1, available at https: / / pubs.broadinstitute.org / mammals / haploreg / haploreg.php, using genotype data from Phase 1 of the 1000G project on the CEU population, where the LD threshold was r 2 >0.8 (Ward LD, Nucleic Acids Res. 2011, 40:D930; 1000 Genomes Project Consortium, Nature, 2015, 536:68).

[0093] Functional annotation of associated variants To predict the most likely functional variants, we mapped variants associated with AE levels against epigenetic marks derived from the Encyclopedia of DNA Elements (ENCODE) and NIH Roadmap Epigenomics project data, including chromatin state (chromHMM) annotations, DNase I hypersensitive regions, transcription factor binding sites, and epigenetic markers (H3K4Me1, H3K4Me3, and H3K27Ac) histone modifications in normal human mammary epithelial cells (HMECs), human mammary fibroblasts (HMFs), BR.MYO (mammary myoepithelial cells), and BR.H35 (mammary vHMECs), as well as in two breast cancer cell lines, MCF-7 and T47D (http: / / genome.ucsc.edu / ENCODE / ). We prioritized variants located in either active promoter or enhancer regions in breast cell lines whose ChIP-Seq data indicated protein binding or whose PWM scores predicted allelic protein binding differences. Two publicly available tools, RegulomeDB and HaploReg v4.1, and the MotifBreakR Bioconductor package, were also used to evaluate these candidate functional variants.

[0094] Breast tumor samples The METABRIC dataset of tumor samples included DNA sequencing data for 2433 samples, and 480 samples were subject to capture-based RNA sequencing studies (Curtis C., Nature 2012,486:346). Sequencing libraries were generated using total RNA generated from frozen tissues by TruSeq mRNA library preparation kit using polyA-enriched RNA (Illumina, San Diego, CA, USA) and enriched with human kinome DNA capture baits (Agilent Technologies, Santa Clara, CA, USA). Six libraries were pooled with 100ng of each library per capture reaction and sequenced (paired-end 51bp) on an Illumina HiSeq2000 platform. A subset of samples with both DNA and RNA sequencing data and harboring PIK3CA mutations was selected for further analysis. The TCGA dataset consisted of 695 TCGA breast cancer samples, and a subset of 289 samples with PIK3CA mutations was selected for further analysis.

[0095] DNA-seq and RNA-seq variant calling in tumors Alignment and pre-processing: Sequence data (FASTQ) mapped to the reference genome (hg19) were aligned using STAR v2.4.1c (https: / / github.com / alexdobin / STAR). A two-pass alignment was performed: splice junctions detected in the first alignment run were used as a guide for the final alignment. Duplicates were marked with Picard v1.131 (http: / / picard.sourceforge.net). Indel realignment and base quality score recalibration were performed using the Genome Analysis Toolkit (GATK). Variant calling and annotation: SNVs and indel variants were called using the GATK Haplotype Caller. Variants were filtered by applying a hard filter using GATK VariantFiltration. Variants were annotated using Ensembl Variant Effect Predictor (VEP). To avoid RNA editing and other RNA-associated variants, and because true allelic imbalance may lead to heterozygous sites being called homozygous in RNA-based genotype calling, heterozygous genotypes were called from the DNA data.

[0096] Statistical analysis of allelic expression imbalance in tumors Prior to analysis, a series of filtering steps were performed to select samples that: 1) passed quality control, 2) had missense mutations, and 3) had a minimum of 30 reads for both RNA and DNA data. Clinical data from METABRIC were updated with the latest available records available from the original studies. Clinical data from TCGA were imported from https: / / portal.gdc.cancer.gov / on November 26, 2018. Three parameters were calculated for all samples: 1) DNA variant allele ratio β = Log2 [(variant allele reads in DNA) / (wild type allele reads in DNA)], which serves to control for sequencing artifacts due to heterozygous genotypes and account for the difference in the frequency of variants in DNA; 2) α = Log2 [(variant allele reads in RNA) / (wild type allele reads in RNA)], which serves as an index of net allelic expression imbalance in tumors; 3) normalized variant allele expression ratio γ = α-β. This is an indication of mutant allele expression imbalance due to cis-regulation alone. Association between allele expression imbalance ratios and clinical data was achieved by bivariate analysis Wilcoxon rank sum test with continuity correction or Kruskal-Wallis rank sum test as shown in tables and figures. P values ​​were adjusted for each study using Bonferroni correction and were considered significant when ≤ 0.05. Correlation analysis of the ratios of "α vs. β" and "α vs. γ" for both sample sets was performed using Pearson test. All statistical analyses and data visualizations were performed using R. Furthermore, we considered mutant allele differential expression (MADE) to be present if |γ| ≥ 0.58 (≥ 1.5-fold difference). In survival analysis, we further divided MADE samples into MADE_wt (γ ≤ -0.58, or preferential expression of wild-type allele) and MADE_mut (γ ≥ 0.58, or preferential expression of mutant allele).

[0097] Survival analysis Kaplan-Meier plots and multivariate Cox proportional hazards models explore the association between allele expression and survival, calculated using the R package "Survival". Death from all causes was used as the end point, and all subjects still alive were censored at the date of last contact. Kaplan-Meier survival curves were compared using the log-rank test. In multivariate analyses, the effect of gamma on overall survival was assessed using Cox proportional hazards models. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated by fitting Cox models while adjusting for age and tumor characteristics, including size, Scarff-Bloom-Richardson histological grade, clinical stage, and estrogen receptor (ER), progesterone (PR), and human epidermal growth factor 2 (Her2) status. All p values ​​were two-sided, and values ​​less than 0.05 were considered statistically significant.

[0098] Availability of code and data The filtered data and the code for variant allele expression imbalance analysis and survival analysis are publicly available at https: / / github.com / maialab / pik3ca.

[0099] Example 1: Normal cis-regulatory mutations affecting the expression of PIK3CA in healthy breast tissue To investigate whether the expression of PIK3CA in normal breast tissue is regulated by cis-regulatory variants, we analyzed data from allelic expression analysis of normal breast tissue from 64 healthy donors. We calculated the logarithm of the ratio of expression of one allele to the other at heterozygous SNP positions (aeSNPs) and calculated differential allelic expression (DAE) when the difference between the alleles was greater than 1.5-fold, i.e., the absolute value of the log2 allele expression ratio ≥ 0.58. This approach robustly detects cis-acting variant effects, as it negates trans-effects that affect both alleles equally. Of the 14 aeSNPs tested, 6 SNPs in PIK3CA showed DAE (daeSNPs) (Figure 1B). All daeSNPs except rs3729679 are in strong linkage disequilibrium (LD) with each other (Table 3).

[0100] The largest proportion of heterozygotes showing DAE (57%) was shown by rs3960984, whereas the smallest portion (14%) was shared by three daeSNPs: rs12488074, rs4855093, and rs9838411. Table 2 summarizes the heterozygote frequency and DAE findings including preferentially expressed alleles for each daeSNP. The range of allele expression imbalance detected was 1.55-fold to 2.14-fold (DAE ratios: 0.63-1.10). For four of the six daeSNPs (rs7636454, rs3960984, rs12488074, rs9838411), the DAE ratios did not show a bilateral distribution, and all samples with DAE preferentially expressed the same allele. For the other two daeSNPs, only one sample each differed in preferentially expressed alleles. These DAE distribution patterns suggest that daeSNPs and possible functional regulatory SNPs (rSNPs) are in strong, though not complete, LD with each other (Xiao R., Genet. Epidemiol. 2011, 35:515).

[0101] Example 2: rSNP candidates in PIK3CA affect the NF-YA binding site in its promoter We then set out to identify rSNPs acting on PIK3CA by performing mapping analysis on all daeSNPs using genotype information from a SNP (NM_006218, GRCh38 / hg38) located within a 250 kb window from the beginning and end of the PIK3CA gene. This work identified a list of 44 candidate SNPs where the DAE observation was significantly evident, along with 229 proxy SNPs (Table 4).

[0102] Functional analysis of these 273 candidates was initiated by in-silico analysis to identify those with high cis-regulatory potential, followed by in-vitro validation, yielding one strong rSNP candidate, rs2699887. This variant was significantly associated with the DAE detected by rs12488074 (permutation p-value = 0.02) (Figure 1C) and was associated with differential total expression of PIK3CA in METABRIC tumors (p-value = 0.012, Figure 4). rs2699887 is located in the first intron of PIK3CA, in a region that is enriched for epigenetic marks and classified as an active promoter in breast mammary epithelial cells (vHMEC) and primary breast myoepithelial cells (Figure 1D). Motif position weight matrix (PWM) analysis of the sequences surrounding rs2699887 suggested that NF-Y protein may have preferential binding affinity to the minor allele (T allele). Electrophoretic mobility shift assays (EMSA) confirmed the preferential protein-DNA binding ability of the minor allele (T allele) of rs2699887, which was confirmed by competition and supershift assays to preferentially bind the transcription factor NF-YA.

[0103] Example 3: Preferential expression of PIK3CA mutant alleles is common in breast tumors DNA copy number changes in tumors are associated with altered gene expression in cis, but whether these differences are also due to cis-regulated expression has not been analyzed before. Given the DAEs identified for PIK3CA in normal breast tissue, PIK3CA somatic mutations were evaluated for functional effects modified by allelic expression imbalance. Preferential expression of gain-of-function mutations may have a stronger clinical impact than low-expressing alleles, which may result in clinical heterogeneity between tumors (Figure 2A). To test this, we performed allelic expression analysis of mutant vs. wild type for breast tumor samples harboring somatic PIK3CA missense mutations from two independent datasets, METABRIC (n=94) and TCGA (n=161). Three allelic ratios derived from DNA-seq and RNA-seq data were evaluated for each mutation: α = log2(number of mutant RNA-seq reads / number of wild-type RNA-seq reads), i.e., net mutant allele expression imbalance; β = log2(number of mutant DNA-seq reads / number of wild-type DNA-seq reads), i.e., the relative copy number of the mutant allele; γ = α-β, the net variant allele expression imbalance normalized for DNA allele copy number imbalance, which corresponds to the putative variant allele expression imbalance due to cis-regulation.

[0104] Alpha indicates global allelic expression imbalance originating from various mechanisms including copy number aberrations, cellular differences, and cis-regulatory mutations; whereas gamma would specifically indicate contributions from cis-regulatory variants (rVar in Figure 2A), including normal genetic mutations, somatic noncoding mutations, and allelic epigenetic changes (Figure 2B). Mutant allelic expression imbalance (alpha ratio) is frequent in breast tumors (METABRIC (75.5%) and TCGA (68%), Table 2). Gamma ratios are similar, 50.4% in METABRIC and 46% in TCGA, indicating that cis-regulatory effects acting on mutations are also frequent in breast tumors. Samples from both cohorts showed large values ​​of preferential allelic expression of mutant alleles (up to 54-fold in METABRIC and 220-fold in TCGA), but not wild-type alleles (fold difference 5.5-fold in METABRIC and 28.2-fold in TCGA) (Table 2, Figure 2B). Similarly, a trend towards more pronounced preferential expression of the variant allele was found, albeit with a smaller fold difference between alleles: 13-fold for METABRIC and 5.5-fold for TCGA for the gamma ratio. Interestingly, samples with differential expression of the variant allele (MADE, defined as |gamma| above a threshold of 0.58) were enriched for preferential expression of the variant allele in both datasets (for METABRIC, Prob(binomial test probability)=0.88, CI95%=[0.75-0.95], p=1.0e-07, and for TCGA, Prob=0.70, CI95%=[0.58-0.80], p=7.6e-04).

[0105] Example 4: Cis-regulatory variants contribute significantly to the imbalance in expression of mutant alleles The contribution of each copy number and cis-regulatory variant to the detected net variant allele expression imbalance was then assessed in the tumors. Global allelic expression correlated with both copy number and cis-regulatory mutations (Figure 2C), although the effect on copy number was approximately twice as large as the effect found for cis-regulatory mutations. The variance in global allelic expression was considered as the sum of the effects of both mechanisms plus a covariance that constituted the expected non-mutual exclusion of both mechanisms acting on any given allele: Var(α)=Var(β)+Var(γ)+2Cov(β,γ).

[0106] The contribution of cis-regulatory variants to the overall allelic expression variance is: [Var(γ)+Cov(β,γ)] / Var(α).

[0107] Cis-regulatory variants accounted for 17.9% and 16.7% of the overall mutant allele expression variability seen in METABRIC and TCGA, respectively. When we assessed how the two mechanisms acted simultaneously in each tumor, the majority of samples (71.28% in METABRIC and 55.28% in TCGA) had positive values ​​of γ and negative values ​​of β (Figure 2D), meaning that the mutant allele was more strongly expressed than the wild-type allele despite its lower genomic content. Only a small proportion of samples showed both higher copies of the mutant allele and preferential allele expression (5.32% in METABRIC and 9.94% in TCGA). When we compared matched β and γ values ​​of PIK3CA, we found no pattern of association between the ratio and cellularity in METABRIC tumors.

[0108] Example 5: MADE defines a malignant subset of PIK3CA mutant tumors To investigate the impact of cis-regulatory differences in PIK3CA mutations on clinical outcomes, univariate survival analysis was performed by classifying the γ ratio into two groups: MADE group (if |γ| ≥ 0.58) and mut=wt group (if |γ| < 0.58). The MADE group had worse overall survival (p-value = 0.03) and disease-specific survival (p-value = 0.012, Figure 3A) than the mut=wt group for METABRIC. The median overall survival was about 8 years for the MADE group and 18.5 years for the mut=wt group (Figure 3A), whereas in disease-specific analysis, 40% of MADE patients died during the follow-up period compared with 20% in the mut=wt group (Table 5). In METABRIC, significantly shorter overall and disease-specific survival was observed for MADE samples that preferentially expressed the wild-type allele (MADE_wt group, p-values ​​= 0.0073 and = 0.0081, respectively) (Figure 3B).

[0109] Example 6: PIK3CA MADE is associated with clinicopathological variables We compared PIK3CA MADE with hormone receptor (ER, PR) and Her2 amplification, which are current standard of care prognostic clinicopathological variables directly and indirectly related to gene expression regulation, respectively (Figure 3C, Table 3). In METABRIC, the α-mean was significantly higher in ER-negative tumors (p-value=0.026), PR-negative tumors (p-value=0.007), and Her2-positive tumors (p-value=0.018) (Figure 5). When evaluating the contribution of cis-regulatory mutations to this association, we also found that higher γ-mean was associated with lower PR expression (p-value=0.040) and Her2-positive tumors (p-value=0.025), but not with ER expression (p-value=0.129) (Figure 3C). This showed that predominance of mutant alleles was associated with poor prognostic variables (ER-negative, PR-negative, Her2-positive), especially in tumors with preferential mutant allele expression. Analysis of a smaller TCGA set confirmed a significant association between PR status and gamma ratio (p-value = 0.003) and between ER status and gamma value (p-value = 0.014) (Figure 3C, Figure 5). These results led to the consideration of MADE in the survival analysis within the ER, PR, and Her2 expression subgroups. No significant differences in overall and disease-specific survival were observed between the MADE and mut=wt groups, but a trend towards worse survival was observed for the MADE group in the Her2-negative subgroup (p-value = 0.056 and 0.027 for overall and disease-specific analyses, respectively). However, further stratification of the MADE group revealed that ER-positive (p-value=0.004), PR-positive (p-value=0.0013), and Her2-negative (p-value=0.0066) tumors identified a poor prognosis in MADE_wt samples (Figure 3D). Thus, MADE defines a malignant subgroup within the normally favorable prognosis ER-positive, PR-positive, and Her2-negative tumor subset. Finally, associations of MADE with other known prognostic variables, including lymph node status, tumor size, grade, and molecular subtypes such as PAM50 and IntClust, did not reveal any additional significant associations.

[0110] overview This finding highlights the importance of assessing the expression levels of mutant oncogenic alleles as part of clinical management. Despite the high frequency of PIK3CA mutations in breast cancer, responses to PI3K inhibitor treatment have not been as successful as hoped, and the prognostic significance of detection of somatic PIK3CA mutations in breast tumors is unclear. First, the examples herein establish the prognostic relevance of transcriptional allelic regulation of somatic mutations in breast cancer and support novel methods of mutation testing in patient management that require consideration of the expression levels of mutations, not just detection at the DNA level. Second, there is a strong need to identify patients who will not respond to specific therapies to prevent unnecessary drug cytotoxicity without therapeutic benefit. The examples herein demonstrate that not all PIK3CA somatic mutations, commonly considered driver events, are equally expressed in tumors, and some barely express the mutations identified at the DNA level. If the targeted somatic mutation is not expressed, we predict that no benefit will be provided by mutation-targeted therapy.

[0111] [Table 1]

[0112] [Table 2]

[0113] [Table 3]

[0114] [Table 4]

[0115] [Table 5]

Claims

1. 1. A method for predicting the prognosis of a cancer patient diagnosed with a tumor characterized by a mutant PIK3CA allele, in particular a breast cancer patient, comprising: The method comprises: a. in a tumor sample obtained from a patient using a method selected from mRNA sequencing, mRNA array, or quantitative real-time PCR: i. the expression level of the mutant PIK3CA allele (ex(mut)), and ii. expression level of the wild type PIK3CA allele (ex(wt)); To determine, b. For cancer patients: i. If ex(mut) is approximately equal to ex(wt), Assigning a probability of favorable prognosis, or ii. ex(mut) significantly exceeds ex(wt), or ex(wt) significantly exceeds ex(mut); assigning a probability of poor prognosis; Including, The method.

2. 1. A method for predicting the prognosis of a cancer patient diagnosed with a tumor characterized by a mutant PIK3CA allele, in particular a breast cancer patient, the method comprising: a. in a tumor sample obtained from a patient using a method selected from mRNA sequencing, mRNA array, or quantitative real-time PCR: i. the expression level of the mutant PIK3CA allele (ex(mut)), and ii. expression level of the wild type PIK3CA allele (ex(wt)); To determine, b. For cancer patients: i. ex(mut) is at least 1.5 times greater than ex(wt), or When ex(wt) is 1.5 times or more than ex(mut), Assigning a probability of poor prognosis; or ii. ex(mut) is less than 1.5 times that of ex(wt); and When ex(wt) is less than 1.5 times ex(mut), assigning a probability of favorable prognosis; Including, The expression level of a mutant PIK3CA allele is determined for at least one PIK3CA single nucleotide polymorphism (SNP) locus, in particular for at least one PIK3CA SNP selected from rs3960984, rs7636454, rs3729679, rs12488074, rs2699887, rs4855093 and / or rs9838411; The method.

3. A poor prognosis is defined as an approximately 45-55% probability of survival after about 11 years, and a good prognosis is defined as an approximately 75-85% probability of survival after about 18 years. The method according to claim 1 or 2.

4. When ex(wt) is 1.5 times or more greater than ex(mut), the cancer patient is assigned a very poor prognosis, defined as an approximately 45-55% probability of survival at about 5 years. The method according to any one of claims 1 to 3.

5. A method for determining the likelihood of response of a tumor disease in a cancer patient, particularly a tumor disease in a breast cancer patient, to treatment with a PI3K inhibitor, the method comprising: a. in a tumor sample obtained from a cancer patient using a method selected from mRNA sequencing, mRNA array, or quantitative real-time PCR: i. Expression level of mutant PIK3CA allele (ex(mut)), ii. Expression level of wild type PIK3CA allele (ex(wt)) To determine, b. For cancer patients: i. assigning a higher probability of response of the tumor disease of the breast cancer patient to treatment with a PI3K inhibitor if ex(mut) is 1.5 times or more greater than ex(wt); or ii. assigning a lower probability of response of the breast cancer patient's tumor disease to treatment with a PI3K inhibitor when ex(mut) is less than 1.5-fold greater than ex(wt), especially when ex(wt) is 1.5-fold greater than ex(mut); The method comprising:

6. The PI3K inhibitor is selected from alpelisib or copanlisib, in particular, the PI3K inhibitor is alpelisib. The method according to claim 5.

7. The expression level of a mutant PIK3CA allele is determined for at least one PIK3CA single nucleotide polymorphism (SNP) locus, in particular: a. rs3960984, b. rs7636454, c. rs3729679, d. rs12488074, e. rs2699887, f. rs4855093, and / or g. rs9838411, For at least one PIK3CA SNP selected from The method according to any one of claims 1 to 6.

8. The expression level of the mutant PIK3CA allele is determined for SNP rs3960984. The method according to any one of claims 1 to 7.

9. The expression level of the mutant PIK3CA allele is determined for a SNP located within an untranslated region of the mutant PIK3CA mRNA transcript. The method according to any one of claims 1 to 8.

10. The expression level of the mutant PIK3CA allele is determined for SNP rs2699887. The method according to any one of claims 1 to 9.

11. The patient has been diagnosed with breast cancer, The method according to any one of claims 1 to 10.

12. The tumor is: a. Estrogen receptor positive or negative; b. Progesterone receptor positive or negative, and / or c. Human epidermal growth factor receptor 2 negative or positive; It has been classified as The method of claim 11.

13. A PI3K inhibitor for use in the treatment of breast patients who have been assigned a higher likelihood of response by the method of any one of claims 5 to 12, in particular a PI3K inhibitor selected from the group comprising alpelisib or copanlisib, in particular a PI3K inhibitor which is alpelisib.

Images