EP3802882A1 - In vitro method for the prediction of response to chemotherapy in triple negative breast cancer patients - Google Patents

Abstract

In vitro method for predicting the response to neoadjuvant chemotherapy (NAC) in triple negative breast cancer patients (TNBC). The present invention relates to medical field, particularly to the oncology field. Precisely, the present invention refers to an in vitro method for the prediction of response to neoadjuvant chemotherapy (NAC), or for selecting a therapy, for triple negative breast cancer patients (TNBC). The method of the invention comprises determining the methylation status of the gene FERD3L and/or TRIP10 in a biological sample obtained from a patient, wherein a higher level of methylation of the gene FERD3L and/or TRIP10, as compared with the level of methylation of at least one of said genes in non-responder patients, is indicative of response to NAC in TNBC patients.

Description

IN VITRO METHOD FOR THE PREDICTION OF RESPONSE TO

CHEMOTHERAPY IN TRIPLE NEGATIVE BREAST CANCER PATIENTS

FIELD OF THE INVENTION

The present invention relates to the medical field, particularly to the oncology field. Precisely, the present invention refers to an in vitro method for the prediction of response to neoadjuvant chemotherapy, preferably based on taxanes and/or anthracyclines (NAC), or for selecting a therapy for triple negative breast cancer patients (TNBC). The method of the invention comprises determining the methylation status of the gene FERD3L and/or TRIP 10 in a biological sample obtained from a TNBC patient, wherein a higher level of methylation of the gene FERD3L and/or TRIP 10, as compared with the level of methylation of at least one of said genes in non-responder patients, is indicative of response to NAC in TNBC patients.

STATE OF THE ART

TNBC refers to any breast cancer that does not express the genes for estrogen receptor (ER), progesterone receptor (PR) and Her2/neu. This makes it more difficult to treat since most hormone therapies target one of the three receptors, so triple-negative cancers often require combination therapies.

Nowadays, chemotherapy is the only proven therapy for triple negative breast cancer subtype and anthracycline/taxane-based regimen has been usually recommended in the neo/adjuvant setting with pathological complete response (pCR) rates between 27-45%. However, newly neoadjuvant chemotherapy (NAC) prospective trials with platinum-based schedules have shown a pCR rate of 54%. A high pCR rate in TNBC is associated with better outcomes while Residual Cancer Burden (RCB) after NAC have a higher relapse risk and poor prognosis. Consequently, increasing the rate of pCR became the end point of neoadjuvant trials with the expectation of translation into improved survival. In this context the identification of predictive factors for able to predict a pCR is nowadays and unmet medical need. In fact, with the initiation of high sequencing technology, several molecular signatures have been developed in the recent years to predict response to NAC. For example, Oncotype-Dx is a diagnostic tool that predicts recurrence risk based on the expression profile of 21 -genes. Oncotype has shown to predict response to NAC mainly in ER-positive breast cancer patients in two different studies. Another assay, called MammaPrint, is a 70-gene expression panel that, as Oncotype, was designed to determine the risk of recurrence in operated early breast cancer. However, in order to predict response to chemotherapy, this assay must be combined with Blue Print, which is an additional 80-genes molecular subtyping assay. Other signatures, such as Endopredict or Prosigna, have also shown predictive power in the neoadjuvant setting mainly in ER+/HER2- breast cancer. More recently, an initial 199-gene signature, E2F4 has shown accurate prediction of response to NAC, even when it was reduced to a 33-gene panel, and has been validated in 1129 patient samples across 5 independent breast cancer neoadjuvant chemotherapy datasets.

However, all these predictive panels are more focused on predicting the response to chemotherapy in estrogen receptor (ER) positive breast. Nevertheless, as cited above, the accurate prediction of response to NAC in TNBC still remains as an unmet medical need.

On the other hand, epigenetic modifications of the DNA such as methylation, due to environmental or external agents, can modulate gene expression with no DNA sequence modification and contribute to disease development in the same way than genetic alterations. Available evidences show that the pattern of DNA methylation of tumor tissue differs from its corresponding normal tissue. In this context, epigenetic changes in tumor DNA before chemotherapy administration could potentially have a predictive role of response to this therapy even more accurate that gene expression profiles.

If fact, the present invention is focused on providing a method for the prediction of response to NAC, preferably to anthracyclines and/or taxanes-based regimens, in TNBC patients, departing from the methylation status of specific genes.

DESCRIPTION OF THE INVENTION

Brief description of the invention

As explained above, the role of an epigenetic methylation-based signature was evaluated in the present invention, in order to provide a method for the prediction of response to NAC (preferably based on taxanes and/or anthracyclines) in TNBC patients. Thus, epigenetic assessment of DNA extracted from archived biopsy TNBC samples previous to NAC was performed. The patients included in the study were categorized according to previous response to NAC in responders to NAC (pCR or RCB=0) or non-responders to NAC (non- pCR or RCB>0). A methyloma study (Infinium HumanMethylation450 array, Illumina) was performed in a discovery cohort. Those methylated genes in the discovery cohort were validated by pyrosequencing (PyroMark Q96 System version 2.0.6, Qiagen) and qPCR in an independent cohort of TNBC patients and in TNBC cell lines. 24 and 30 patients were included in the discovery and validation cohorts respectively. In the discovery cohort, 9 genes were differentially methylated: 6 showed higher methylation status in non-responder patients ( LOC641519 , LEF1, HOXA5, EVC2, TLX3 and CDKL2 ) and 3 greater methylation status in responders patients ( FERD3L , CHL1 and TRIP10 ). After the validation, a 2-genes (FER3L and TRIP 10) epigenetic score predicted RCB=0 with an area under the curve (AUC) ROC 0.9056 (CI95%=0.805-1.000). Patients with a positive epigenetic 2-genes score had 78.6% RCB=0 vs. only 10.7% RCB=0 if signature was negative.

The statistical analysis of the method of the invention based on determining the methylation state of both genes, gives rise to a ROC curve showing an AUC of approximately 0.90 (see Figure 4 and Example 19) . However, the statistical analysis based on the determination of the methylation state of just one of the two genes reveals an AUC of 0.8115 (CI95%=0.680- 0.943) for FERD3L (see Figure 5) and AUC of 0.8032 (CI95%=0.670-0.937) for TRIP 10 (see Figure 6). This means that the method of the invention for predicting the response to NAC in TNBC patients can be based on the determination of the methylation status of just one of the genes FERD3L or TRIP 10, preferably FERD3L. Moreover, this data show that the combination of both genes considerably improves the prediction capacity of the individual genes, and thus the method of the invention for predicting the response to NAC in TNBC patients based on the determination of the methylation status of at least the two genes FERD3L or TRIP 10, is particularly preferred.

Consequently, the present invention suggests a role of the methylation of FERD3L and/or TRIP 10 in the prediction of response to NAC treatment in TNBC. The gene FERD3L (Fer3 like bHLH transcription factor), also named NAT03 or N-TWIST, is a gene located on chromosome 7 and is a basic helix-loop-helix (bHLH) transcription factor. The UCSC Refgene accession number (Refseq) of FERD3L is Ref Seq: NM 152898.

TRIP 10 gene (thyroid hormone receptor interactor 10), also known as CIP4, is located in chromosome 19 and belongs to the minor histocompatibility antigens family. It is a protein that is involved in diverse signaling pathways and it has diverse functions in wide variety of cell types. The UCSC Ref gene accession number (Refseq) of TRIP10 is Ref Seq: NM_004240.

The present invention suggests that response to NAC treatment can be predicted accurately with an epigenetic signature of the methylation status of FERD3L and/or TRIP 10 in patients with TNBC. Hypermethylation of both FERD3L and TRIP 10 predicts a 78.6% pCR (RCB=0), which almost doubles the predictive potential of the breast cancer subtype (TNBC has approximately a 30-40% pCR after NAC) [Carey LA, Dees EC, Sawyer L, et al. The triple negative paradox: primary tumor chemosensitivity of breast cancer subtypes. Clin Cancer Res 2007;13:2329-34]

Thus, the first embodiment of the present invention refers to an in vitro method for the prediction of response to NAC in TNBC patients, which comprises determining the methylation status or the expression level of the gene FERD3L in a biological sample obtained from the patient, wherein a higher level of methylation or a lower expression level of the gene FERD3L, as compared with a reference level of methylation or the expression level of the gene FERD3L measured in non-responder patients, is indicative of response to neoadjuvant chemotherapy in TNBC patients. In a preferred embodiment, the gene FERD3L is methylated in its promoter region.

The second embodiment of the present invention refers to an in vitro method for selecting a therapy for TNBC patients, which comprises determining the methylation status or the expression level of the gene FERD3L in a biological sample obtained from the patient, wherein a higher level of methylation or a lower expression level of the gene FERD3L, as compared with a reference level of methylation or the expression level of the gene FERD3L measured in non-responder patients, is indicative of response to NAC in TNBC patients. In a preferred embodiment, the gene FERD3L is methylated in its promoter region. In a preferred embodiment, the above cited methods comprise determining the methylation status or the expression level of at least the two genes FERD3L and TRIP10 in a biological sample obtained from the patient, wherein a higher level of methylation or a lower expression level of at least the two genes FERD3L and TRIP 10, as compared with a reference level of methylation or of expression level of the genes FERD3L and TRIP10 measured in non responder patients, is indicative of response to NAC in TNBC patients. In a preferred embodiment, the genes FERD3L and TRIP 10 are methylated in their promoter region.

The third embodiment of the present invention refers to the in vitro use of the methylation status of the gene FERD3L for the prediction of response to NAC in TNBC patients. In a preferred embodiment, the gene FERD3L is methylated in its promoter region

The fourth embodiment of the present invention refers to the in vitro use of the methylation status of the gene FERD3L for selecting NAC for TNBC patients. In a preferred embodiment, the gene FERD3L is methylated in its promoter region

In a preferred embodiment, the present invention refers to the in vitro use of the methylation status of the genes FERD3L and TRIP 10, for the prediction of response to NAC in TNBC patients or for selecting NAC as a treatment for TNBC patients. In a preferred embodiment, the genes FERD3L and TRIP 10 are methylated in their promoter region.

In a preferred embodiment, the neoadjuvant chemotherapy is non-platinum neoadjuvant chemotherapy comprising taxanes and/or anthracyclines; or platinum-based chemotherapy, for example taxanes and carboplatin.

In a preferred embodiment, the gene FERD3L is methylated in the promoter region CpG cgl0043037. It is located in the following position of the genome chr7: 19185407. In a more preferred embodiment, the gene FERD3L is methylated in a promoter region CpG which is up to 100 base pairs upstream or downstream from the CpG cgl0043037, preferably between 1 and 80 nucleotides, and more preferably between 1 and 60 nucleotides, upstream or downstream from the CpG cgl0043037. In a preferred embodiment, the gene TRIP 10 is methylated in the promoter region CpG cg02085507. It is located in the following position of the genome: Chrl9:6739l92. In a more preferred embodiment the gene TRIP 10 is methylated in a promoter region CpG which is up to 100 base pairs upstream or downstream from the CpG cg02085507, preferably between 1 and 80 nucleotides, and more preferably between 1 and 60 nucleotides, upstream or downstream from the CpG cg02085507.

Thus, in a preferred embodiment of the invention the gene FERD3L and TRIP 10 are methylated in a region as shown in Table 1, wherein it is stated the CpGs studied by pyrosequencing in the discovery cohort (DC) and validation cohort (VC) in order to validate the methylation status of the candidate genes FERD3L and TRIP 10 which is identified in the 450k array (Illumina). In bold, CpGs from 450k array are represented in bold. Moreover, this Table 1 also shows consecutive CpGs represented in normal type which are methylated in TNBC patients responding to NAC. In a preferred embodiment, the method of the invention is performed by determining the methylation status of any of the target CpGs of the Table 1, preferably in combination with any of the consecutive CpGs also disclosed in this Table 1:

Table 1

The fifth embodiment of the present invention refers to a method for treating patients suffering from TNBC with NAC which comprises, as a step previous to the treatment with NAC, predicting the response to NAC by means of any of the above described methods which comprise determining the methylation status of the gene FERD3L in a biological sample obtained from the patient, wherein the patient will be treated with NAC if a higher level of methylation of the gene FERD3L , as compared with a reference level of methylation of the gene FERD3L measured in non-responder patients, is finally identified.

The sixth embodiment of the present invention refers to a kit comprising reagents for the determination of the methylation status of at least the genes FERD3L and TRIP 10, or to the use of this kit for the prediction of response to neoadjuvant chemotherapy in triple negative breast cancer patients or for selecting neoadjuvant chemotherapy as a treatment for triple negative breast cancer patients.

For the purpose of the present invention the following definitions are given:

• As used herein, the term“triple negative breast cancer” or“TNBC” refers to a breast cancer that is characterized as being estrogen receptor (ER) negative, progesterone receptor (PR) negative and human epidermal growth factor receptor 2 (HER-2) negative. Thus, the level of expression of each one of ER, PR and HER-2 may be reduced when compared to a non-cancerous sample, or an ER+ve, PR+ve and HER2 +ve cancerous sample, or which is characterized by a level of expression of each one of ER, PR and HER-2 which is not significantly different from the level of expression of a housekeeping gene, or which is characterized by the absence of a detectable level of expression of each one of ER, PR and HER-2, or which is characterized by the absence of expression of each one of ER, PR and HER-2. Alternatively, the characterization of the“triple negative breast cancer” or“TNBC” can be carried out by following anatomic pathology criteria according to the procedure described in Example 3.

• The expression“neoadjuvant chemotherapy” refers to the chemotherapy given as a first step to shrink a tumor before the main treatment, which is usually surgery. In a preferred embodiment, “neoadjuvant chemotherapy” refers to non-platinum neoadjuvant chemotherapy comprising taxanes and/or anthracyclines; or platinum- based chemotherapy, for example taxanes and carboplatin.

• The term“sample” or "tumor sample" refers to a sample comprising tumor material obtained from a cancer patient. The term encompasses tumor tissue samples, for example, tissue obtained by surgical resection and tissue obtained by biopsy, such as for example, a core biopsy or a fine needle biopsy. In a particular embodiment, the tumor sample is a fixed, wax-embedded tissue sample, such as a formalin- fixed, paraffin-embedded tissue sample. Additionally, the term "tumor sample" encompasses a sample comprising tumor cells or circulating tumor nucleic acid obtained from sites other than the primary tumor, e.g., circulating tumor cells. Thus, according to the present invention, the sample can be a liquid biopsy, for example blood, serum or plasma.

• As used herein, the term "methylation" will be understood to mean the presence of a methyl group added by the action of a DNA methyl transferase enzyme to a cytosine base or bases in a region of nucleic acid e.g. genomic DNA.

• Accordingly, the term, “methylation status” as used herein refers to the level of methylation measured in the patient which is being analyzed to be subjected to NAC. The“methylation status” may be a higher/lower level of methylation as compared with the reference or control level.

• The expression“higher level of methylation” refers to an statistically significant increase in the relative amount of methylation of a nucleic acid e.g., genomic DNA, as compared with the amount of methylation measured in a patient used as control/reference that, in this case, are patients which do not respond to the treatment. Thus, in the present disclosure, the“higher level of methylation” is determined with reference to a baseline level represented by the methylation status of a given genomic region in a sample obtained from non-responder patients. For example,“higher level of methylation” may be at least 2% greater than the baseline level of methylation, for example at least 5% greater than the baseline level of methylation, or at least 10% greater than the baseline level of methylation, or at least 15% greater than the baseline level of methylation, or at least 20% greater than the baseline level of methylation, or at least 25% greater than the baseline level of methylation, or at least 30% greater than the baseline level of methylation, or at least 40% g greater than the baseline level of methylation, or at least 50% greater than the baseline level of methylation, or at least 60% greater than the baseline level of methylation, or at least 70% greater than the baseline level of methylation, or at least 80% greater than the baseline level of methylation, or at least 90% greater than the baseline level of methylation. • As used herein, a“control, reference or baseline level of methylation” shall be understood to mean a level of methylation detected in a corresponding nucleic acid from a non-responder patient. Thus, the patient is likely to respond to NAC, with a given sensitivity and specificity, if a“higher level of methylation” is measured in this patient as compared with a“control, reference or baseline level of methylation” measured in non-responder patients.

• As used herein, a“CpG dinucleotide”,“CpG methylation site” or equivalent, shall be taken to denote a cytosine linked to a guanine by a phosphodiester bond. CpG dinucleotides are targets for methylation of the cytosine residue and may reside within coding or non-coding nucleic acids. Non-coding nucleic acids are understood in the art to include introns, 5'- untranslated regions, 3' untranslated regions, promoter regions of a genomic gene, or intergenic regions.

• The expression “target CpG” refers to the specific CpG which is found to be differentially methylated in the present invention. In a preferred embodiment the “target CpG” is Cgl0043037 (promoter region of FERDL3 ) or Cg02085507 (promoter region of TRIP 10).

• The expression“consecutive CpG” refers to other CpGs which may be differentially methylated and they are in region, determined by bisulphite pyrosequencing, which is up to 100 base pairs upstream or downstream from the“target CpG”, preferably between 1 and 80 nucleotides, and more preferably between 1 and 60 nucleotides, upstream or downstream from the“target CpG”.

• As used in the present invention“determining the methylation status of the gene FERD3L or TRIP 10” means that the method of the invention is performed by measuring the methylation status of any region of the gene, also comprising any regulatory sequence which is capable of increasing or decreasing the expression of said genes. Particularly, said regulatory sequence includes the promoter which is a region of DNA that initiates transcription of the gene. Promoters are located near the transcription start sites of genes, on the same strand, upstream or downstream on the DNA (towards the 5' region of the sense strand).

• As used herein, the term“patient” shall be taken to mean any animal including a human, preferably a mammal. Exemplary subjects include but are not limited to humans, primates, livestock (e.g. sheep, cows, horses, donkeys, pigs), companion animals (e.g. dogs, cats), laboratory test animals (e.g. mice, rabbits, rats, guinea pigs, hamsters), captive wild animals (e.g. fox, deer). Preferably the mammal is a human or primate. More preferably the mammal is a human.

• The term "comprising" it is meant including, but not limited to, whatever follows the word "comprising". Thus, use of the term "comprising" indicates that the listed elements are required or mandatory, but that other elements are optional and may or may not be present.

• By "consisting of’ is meant including, and limited to, whatever follows the phrase “consisting of’. Thus, the phrase "consisting of’ indicates that the listed elements are required or mandatory, and that no other elements may be present.

Description of the figures

Figure 1. Genes FERD3L (A) and TRIP 10 (B) show a consistent DNA methylation profile of consecutive CpGs from Illumina 450k array. In a preferred embodiment, the gene FERD3L is methylated in the promoter region CpG eg 10043037 and the gene TRIP 10 is methylated in the promoter region CpG cg02085507.

Figure 2. It shows the validation results of the candidate genes by pyrosequencing. Data were replicated in FERD3L (A) and TRIP 10 (B) genes. It shows that the methylation level of both genes are statistically higher in TNBC patients responding to NAC as compared with non responder patients (p<0.05). (C) Validation results by pyrosequencing for FERD3L obtained in 450k array in the Validation Cohort (VC, N=30). Methylation data were only replicated for FERD3L gene. Values were statistically different when compared Non-Responders vs Responders group showing low methylation in non-responders patients (p= 0.0087).

Figure 3. (A) FERD3L methylation levels (pyrosequencing) and gene expression levels

(qRT-PCR) in a set of TNBC cell lines. FERD3L was methylated in all the cell lines studied (>40%) correlating with the gene expression detected. (B) Variation in FERD3L methylation and gene expression in MDA-MB-436 cell line after AZA (5uM) treatment. Cells showed a decrease in methylation level (p= 0.05) that correlated with a statistically significant increase in gene expression (p= 0.0022) when they were treated with AZA and compared with control cells. (C) qPCR results for FERD3L gene in TNBC patients (DC+VC). Non-Responders group showed high gene expression levels than Responders group (p= 0.04) correlating with methylation levels obtained both in 450k array and pyrosequencing. (D) Correlation between methylation and gene expression for FERD3L gene in a population of breast cancer patients (N=7l3) from TCGA database. Result showed a correlation between a high gene methylation and a low gene expression when analyzed all the CpGs in the FERD3L gene promoter or only the cgl0043037 identified in 450k array.

Figure 4. ROC curve with AUC 0.9056 (CI95%=0.805-1.000) finally obtained with a model wherein the methylation status of both genes FERD3L and TRIP 10 was measured.

Figure 5. ROC curve with AUC of 0.8115 (CI95%=0.680-0.943) finally obtained with a model wherein the methylation status of FERD3L gene was measured.

Figure 6. ROC curve with AUC of 0.8032 (CI95%=0.670-0.937) finally obtained with a model wherein the methylation status of TRIP 10 gene was measured.

Detailed description of the invention

Example 1. Patients and clinical treatment.

Patients with TNBC were treated with NAC following a dosage regime commonly used in the general practice in this field, preferably anthracyclines (for example Epirubicin 75-100 mg/m2 and/or Doxorubicin 60 mg/m2) and/or taxanes (for example Paclitaxel 80-100 mg/m2 and/or Docetaxel 75 mg/m2), in the Hospital Clinico of Valencia and diagnosed between the years 2005 and 2015 were selected from a prospectively maintained database. Clinical inclusion/exclusion criteria were: 1) Age>l8 years, 2) performance status 0-2 at the moment of initiating NAC, 3) histological diagnosis of an invasive triple negative carcinoma of the breast, after confirmation of ER, PR and HER2 negativity by immunohistochemical techniques according to the pathological guidelines, 4) stage I-III, 5) patients considered candidate for NAC and receiving taxanes and/or anthracy cline for at least 3 cycles or 3 months if weekly regimens in the neoadjuvant setting was given, 6) availability of tumor sample archived from biopsy, 6) availability of specimen from surgery including RCB index assessment, 7) appropriate renal and hepatic function for receiving NAC (as considered by the clinician). From an initial analysis of the database 70 patients were identified. After assessment of pretreatment samples, only those patients that had a tumor percentage greater than 40% and reached 500 ng in the quantification after DNA extraction were used. Out of a total of 70 patients initially selected, 54 patients met the established criteria. Of these, a group of 24 patients (10 responders to NAC or RCB=0 and 14 non-responders to NAC or RCB>0) were selected for the Discovery Cohort (DC). The remaining 30 patients (9 patients RCB=0 and 21 RCB>0), were included in the Validation Cohort (VC). The type of NAC, clinical assessment, hematologic and non-hematologic toxicity monitoring and imaging response assessment was performed as part of the clinical routine practice and decided by the corresponding physician. Archived samples from those patients accomplishing criteria were obtained. This study was approved by the local clinical Research Ethics Committee. Patients were requested to sign informed consent for tumor tissue molecular analysis and biobanking.

Example 2. Sample size calculation.

Sample size calculation was based on previous data from a recent series reporting that the proportion of TNBC patients treated with NAC with anthracyclines and taxanes obtaining a RCB=0 was 29.7%. Thus it could be stated that if the methylation data would be able to predict two groups of responders vs non-responders, the proportion of patients with RCB 0 (responder group) could be approximately 45% while the proportion of RCB 0 in the non responder group would be 10%. With an alfa error of 5% and a power of 80% the sample size needed to identify this difference between both proportions will be 44 patients (22 patients in each group). A 15% drop-out was foreseen. According to this the final sample size should be 52 patients. Thus, 24 patients were included in the discovery cohort. In this cohort a first analysis of the methylation changes was performed and those genes with highest significance were identified. Thereafter, 30 more patients were included in a validation cohort in which those selected genes were validated. With the sum of the patient of both cohorts, a final population of 54 patients was obtained on which the final analysis was performed.

Example 3. Tumor samples from TNBC.

In all cases, tumor samples were obtained from the biopsy performed before exposure to any systemic anticancer treatment. Pretreatment samples were obtained by interventional radiologists using ultrasound-guided core needle biopsy. One of the cores was placed on OCT and stored at -80°C, and freeze-fresh cuts were done to check the tumor percentage after hematoxylin-eosin (H&E) staining. In tumor-rich samples, newer cuts were done to RNA/DNA extraction for the methylation study. The rest of the cores, as well as surgery specimens, were fixed for 8-24 hours in 10% neutrally-buffered formaldehyde and embedded in paraffin. The samples were analyzed by an experienced pathologist (OB) who performed histological diagnosis of invasive breast carcinoma and the determination of ER, PR, HER2 and Ki67. For each case, three-micrometer sections from formalin-fixed, paraffin-embedded tumor tissue were set for immunohistochemistry (IHC) together with appropriate positive controls. Until then, primary antibodies against ER (clone 6F11) and PR (clone 1 A6) were purchased from Novocastra (Newcastle Upon Tyne, UK). Dako Autostainer Link48 was used, with Dako EnVision FlexTM as developing technique. In the last years, monoclonal rabbit antibodies against ER (clone SP1), PR (clone 1E2) and Ki67 (clone 30-9) from Ventana (Tucson, AZ) were employed, using the Ventana Benchmark XT automated platform with the Ventana Ultraview DAB detection kit. IHC determination of HER2 changed in 2009 from semi-automated (HerceptestTM kit, Dako; Glostrup, Denmark) performance to fully automated work (PathwayTM antiHER2, clone 4B5; Ventana; Tucson, AZ) using Benchmark XT platform. In the diagnostic routine conditions ER and PR were assessed by determining the percentage of positive nuclear staining. Tumors with a staining of at least 1% preferably of at least 10% of nuclei are considered ER or PR negative. For the assessment of HER2, ASCO/CAP recommendations were used. Diagnosis of TNBC was done according to IHC results (ER, PR and HER2 -negative tumors). Cells with nuclear staining for Ki67 antibody were counted and expressed as a percentage.

Example 4. RCB calculation.

The pathological response after NAC was evaluated in the surgery specimens (either lumpectomy or mastectomy; both with additional sentinel lymph node biopsy and/or axillary clearance) by the study pathologist (OB). Tissue samples were fixed in 10% neutrally buffered formalin and paraffin embedded following standard protocols of the histology laboratory. Three-micrometer sections were stained with H&E for histologic review. pCR after NAC was diagnosed in case of total absence of invasive residual tumor both in breast tissue and axillary nodes. In cases with residual tumor, the Symmans method was used to assess the Residual Cancer Burden (RCB) [Symmans WF, Wei C, Gould R, Yu X, Zhang Y, Liu M, Walls A, Bousamra A, Ramineni M, Sinn B, Hunt K, Buchholz TA, Valero V, Buzdar AU, Yang W, Brewster AM, Moulder S, Pusztai L, Hatzis C, Hortobagyi GN. Long-Term Prognostic Risk After Neoadjuvant Chemotherapy Associated With Residual Cancer Burden and Breast Cancer Subtype. J Clin Oncol. 2017 Apr l;35(l0): 1049-1060. doi: 10.1200/JC0.2015.63.1010. Epub 2017 Jan 30] RCB is a factor calculated from different variables measured in both the primary tumor bed and the lymph nodes in the surgery specimen after NAC. Variables measured were the maximum diameters of the residual invasive tumor (in millimeters), percentage of tumor cellularity in tumor bed, percentage of in situ tumor in this bed, number of positive lymph nodes and diameter of the greatest metastasis (in millimeters). Using the Residual Cancer Burden Calculator of MD Anderson Cancer Center available on Internet

(http://www3. mdanderson.org/app/medcalc/index. cfm?pagename=jsconvert3), a numeric value (RCB Index) and also a RCB category were obtained in each case. A value of RCB=0 implies a complete pathological response, whereas values of RCB > 0 will indicate that there is still residual tumor. In these cases, three categories were assigned by the RCB calculator (RCB-I, II or III). This system has been validated as a useful tool to predict the clinical outcome after NAC, with significant prognostic value. It has been recently recommended by international working groups to assess the pathological response after NAC in breast cancer clinical trials, and it is also the recommended system to determine the breast cancer response to NAC in the clinical setting in Spain.

Example 5. TNBC cellular lines.

For preclinical analysis, the following commercial TNBC cell lines were used: HCC1937, HCC-1143, HCC-38, MDA-MB-231 and MDA-MB-436. Each cell line was grown following the directions specified (American Type Culture Collection, ATCC). Culture conditions were 5% C0₂ and 37°C atmosphere, DMEM F12 medium for MDA-MB-231 cells, RPMI 1640 Glutamax medium for HCC-1937, HCC-1143 and HCC-38 cells, and DMEM Glutamax medium for MDA-MB-436 cells, all supplemented with FBS 10% and Penicillin/Streptomycin 1%.

Example 6. DNA extraction and bisulfite conversion.

In all cases, samples were considered valuable for DNA extraction if tumor percentage in tumor samples was >25%. Those cases with a lower percentage of tumor cells were discarded. Thereafter, DNA extraction was performed using the kit "DNA PPPE QIAamp Tissue" (Qiagen) according to the manufacturer’s instructions. For the DC, methylome study was performed with Illumina technique. In this case, DNA extraction was performed by DNA purification protocol with NaCl. For the VC methylome was studied with pyrosequencing. In this cohort both OCT or paraffin embedded TNBC tissue were used depending on availability. DNA from TNBC cells lines (control and treated with AZA) was extracted for methylation study by pyrosequencing. When at 90% of confluence, cells pellets were isolated for DNA extraction using Trizol reagent. DNA samples were quantified using PicoGreen method (Invitrogen) and quality was measured based on A260/280 and A260/230 ratios (Nanodrop, Thermo Scientific). After that, an electrophoresis (1.5 % agarose gel) were done to discard DNA fragmentation or RNA contamination. Methylome study was performed with EZ-96 DNA Methylation kit following the manufacturer’s recommendations (Zymo Research Corp.). According to these recommendations at least 500 ng of DNA were needed in order to obtain an effective DNA bisulfite conversion to transform non-methylated cytosines to uracil (methylated cytosines remain unchanged).

Example 7. DNA methylation analysis by Infinium Human Methylation 450 BeadChip array.

In order to analyze genome-wide methylation in the discovery cohort, microarray-based DNA methylation analysis was conducted with the Infinium Human Methylation 450 BeadChip (450k array; Illumina, San Diego, CA), that covers >450,000 CpG sites targeting nearly all of the RefSeq genes (>99%) including coding and non-coding genes without bias against those lacking CpG islands. The design further aimed to cover not only promoter regulatory regions but also CpGs across gene regions to include the 5 '-untranslated regions (5' UTRs), the first exons, the gene bodies and the 3 '-untranslated regions (3' UTRs) with an average of 17 CpG sites per gene. DNA quality checks, bisulfite modification, hybridization, data normalization, statistical filtering, and b value calculations were performed as described elsewhere. A total of 500 ng of tumor DNA samples were selected for bisulfite conversion (Zymo Research; EZ-96 DNA Methylation™ Kit) and hybridization to the Infinium Human Methylation 450 BeadChips (Illumina) following the Illumina Infinium HD methylation protocol. The DNA concentration of the quality control sample standards was measured using the PicoGreen method (Invitrogen). Analysis with 1.3% agarose gel electrophoresis permitted the exclusion of samples with possible DNA fragmentation or RNA contamination. Whole-genome amplification and hybridization were then performed on the BeadChip and followed by single-base extension and analysis on a HiScan SQ module (Illumina) to assess the cytosine methylation states. The annotation of the CG Islands (CGIs) used the following categorization: 1) shore, for each of the 2-kb sequences flanking a CGI; 2) shelf, for each of the 2-kb sequences next to a shore; and 3) open sea, for DNA not included in any of the previous sequences or in CGIs. The transcription start site 200 and the transcription start site 1500 indicate the regions either 200 or 1500 bp from the transcription start site, respectively.

Methylation score of each CpG was represented as beta (b) value and were previously normalized for color bias adjustment, background level adjustment and quantile normalization across arrays. Probes and sample filtering involved a two-step process for removing SNPs and unreliable betas with high detection P value P > 0.001. Sex chromosome probes were also removed. After this filtering, the remaining CpGs were considered valid for the study. Differentially methylated CpG sites (DMCpGs) between the responder and non-responder groups were identified: for each probe/CpG, the sets of methylation b values belonging to the responders (R) and non-responders (NR) group were compared to obtain 1) DMCpGs with a significant P value < 0.05 (1030 CpGs); 2) DMCpGs with average b values between R and NR groups higher than 0.20 (133 out of 1030 CpGs); 3) DMCpGs located in islands or shores of promoter regions of (candidate) genes (35 out of 133 CpGs); 4) DMCpGs with intragroup standard deviation less than 0.20 (11 out of 35 CpGs). This final filter with 11 CpGs belonging to 11 genes yielded the best candidates for validation.

Example 8. Gene Ontology (GO) enrichment analysis.

A gene ontology (GO) analysis was performed to estimate the enrichment of the DMCpGs identified in particular biological processes. This analysis detects the significant over representation of GO terms in one of the sets (i.e. list of selected genes) with respect to the other for the entire genome. GO terms with an adjusted P value < 0.05 were considered significant.

Example 9. DNA methylation analysis by bisulfite pyrosequencing.

Pyrosequencing was performed as a validation of the methylation data obtained by Illumina in the DC and in an independent group of TNBC patients (VC). Moreover, pyrosequencing was used for studying the methylome in TNBC cell lines before and after exposure to AZA. Quantitative DNA methylation analysis was performed by bisulfite pyrosequencing of consecutive cytosines located in islands or shores of promoter regions of candidate genes. Bisulfite conversion of 500 ng of each DNA sample was performed with EZ DNA Methylation-Gold Kit (Zymo Research) according to the manufacturer’s recommendations. Primer sequences (see Table 2 wherein it is disclosed the primers sequences used for pyrosequencing during the validation assay of the candidate genes obtained from 450k array) were designed with PyroMark Assay Design 2.0 (Qiagen). PCRs for gene promoters were performed with 1 mΐ of bisulfite converted DNA with biotinylated primers using an annealing temperature of 60°C and 50 cycles. PCR products were verified on 2% agarose gels before pyrosequencing analysis. Pyrosequencing was performed using a Pyro Gold SQA™ Reagent Kit (Qiagen) in a PyroMark Q96 System version 2.0.6 (Qiagen) according to the manufacturer’s instructions. CpG site methylation quantification was obtained using Pyro Q- CpG 1.0.9 (Qiagen).

Table 2

Example 10. Gene expression studies and correlation with methylation levels. Total RNA was isolated from OCT/ FFPE TNBC tumor samples by mirVana Isolation Kit (Ambion) and from TNBC cell lines using TRIzol (Invitrogen) according to the manufacturer's protocol, treated with DNase I for 30 min at 37°C followed by phenol: chloroform extraction and ethanol precipitation at -80°C. Next, 1-2 pg of RNA were retrotranscribed using the Thermo ScriptTM RT-PCR System (Invitrogen) according to the manufacturer's recommendations. Quantitative RT-PCR (qRT-PCR) reactions were performed in triplicate on an Applied Biosystems 7,900HT Fast Real-Time PCR system using 25-50 ng cDNA, TaqMan expression assays: 0.5 pL of specific TaqMan probes (Applied Biosystems) and 5 pL of Universal PCR Master Mix, no AmpErase® UNG (Applied Biosystems) in a final volume of 10 pL. Gene expression was finally assayed with GAPDH as endogenous control and using the delta delta Ct method.

Example 11. Demethylation studies in TNBC cell lines.

TNBC cell lines were treated with AZA demethylating agent (5-aza-2'-deoxycytidine) in order to reverse the DNA methylation status and evaluated restatement of genes expression by qPCR. MDA-MB- 436 cells (ATCC Manassas, VA, USA) were seeded in 6-wells-plates and maintained in DMEM GlutaMax medium with 10% fetal bovine serum (FBS) and 1% antibiotics (100 U/mL penicillin and 100 mg/L streptomycin) in a humidified atmosphere with 5% C0₂ at 37°C. When reached 60% of confluence, 24h after seeding, cells were treated with AZA agent (5uM) for 72 hours changing the medium each day. Cellular pellets were isolated for RNA extraction and perform gene expression assays.

Example 12. Methylation and gene expression analysis in TCGA database.

DNA methylation and expression data from patients with invasive breast carcinoma (BRCA) were obtained from public repository The Cancer Genome Atlas (TCGA) using the Methhc database. To analyze the association between the DNA methylation and the expression of the gene FERD3L , we used paired DNA methylation and expression data from 713 patients with BRCA obtained from Infinium 450K array and RNA-Seq, respectively. For DNA methylation of FERD3L we considered the average of the methylation data of the promoter region or the methylation value of an individual CpG (cgl0043037) located at the promoter region of FERD3L and detected by the Infinium 45 OK array. Example 13. Statistical analysis.

Discovery study: Data were summarized by mean, standard deviation (SD) or median. Differences in methylation and expression values among groups were assessed using the nonparametric Wilcoxon rank sum test. Receiver Operating Characteristic (ROC) curves was used to assess diagnostic the predictive capacity of the candidate biomarkers. Area under the curve (AUC) was computed for each ROC curve, and 95% confidence intervals (Cl) were also estimated by bootstrapping with 1,000 iterations. Sensitivity and specificity were estimated at the optimal cut-off point according to Youden criterion. Globally, a two-tailed P value of less than 0.05 was considered to indicate statistical significance. All statistical analyses were performed using GraphPad Prism 7 and R software (version 3.2.0). For the analysis of the results obtained from the validation study and gene expression study, we analyzed the significance of any methylation or gene expression difference between the responders and non-responders patients using the non-parametric Mann Whitney test for independent samples. For gene expression studies in cell lines, the AZA samples and control samples were compared using a two-tailed Student t-test. All data presented includes the standard deviation (SD) considering significant values (p) lower than 0.05. The genes that remained different in the validation cohort with a signification level of p<0.2 were selected for confirmation in the whole cohort of 54 patients including both cohorts. These data obtained from the methylation levels of the selected genes were analyzed with binary logistic regression. For the selection of variables the Akaike information criterion (AIC) was used. The models were assessed by construction of ROC curves.

Example 14. Clinical characteristics of TNBC patients.

Fifty-four patients were included: 24 in the DC and 30 in the VC. After biopsy, all patients were treated with NAC based on a taxane and/or anthracycline regimen. The median patient age was 47.9 years (range 27.2-78.9) and all of them were considered TNBC according to immunohistochemistry for ER, PR and HER2. Patients were classified in Responders (R) if RCB=0 or Non-Responders (NR) if RCB > 0.

Example 15. Analysis of DNA methylation in responder vs non responder TNBC patients by Illumina. Discovery cohort. A genome-wide methylation study was performed in the DC (N=24, 10 responders patients (RCB=0) and 14 non-responders (RCB > 0)). The study was done using a microarray-based DNA methylation by Illumina. The analysis of the methylation data showed 133 CpGs sites (71 genes) with differences in methylation levels > 20% (p<0.05) that clearly distinguished responders patients (treatment sensitive) from non-responder patients (treatment resistant). Some of these genes were involved in biological functions such as DNA repair, cell adhesion, and transcription regulation among others. Thirty-five CpGs located in promoters, islands or shores from 23 genes were selected from the 133 initial CpGs to further validations. Of these, taking into account a standard deviation intragroup <0.2, only 11 CpGs corresponding to 11 genes showed significant methylation differences, and only 10 CpGs from 9 genes showed a consistent methylation profile. These candidate genes were LOC641519, LEF1, HOXA5, EVC2, TLX3 and CDKL2 with high methylation in Non- Responders group, and genes FERD3L, CHL1 and TRIP 10 with high methylation in Responders group. See Figure 1 wherein genes FERD3L and TRIP10 show a consistent DNA methylation profile of consecutive CpGs. Two groups are differentiated: Genes with high methylation in Non-Responders (Resistance) and genes with high methylation in Responders (Sensitive). Table 3 shows eleven CpGs differentially methylated, corresponding to 11 genes, showed significant methylation differences between non-responders and responders patients: 6 genes (LOC64l5l8;LEFl; HOXA5; EVC2; CDKL2; TLX3) presented a methylation increase in non-responders group vs responders, and 5 genes (ZFHX4;LOClOOl92378; FERD3L ; CHL1; TRIP 10) decreased methylation level in non-responders patients compared to those who respond to NAC treatment.

Table 3

IELATtON_TO_

UCSC.CPGJSIA

_

«0708550719 6739193 TWPIO TSS 1.10C M Jihore -0 77 Example 16. Methylation gene validation by pyrosequencing.

Validation Cohort. A pyrosequencing study in the same cohort (DC) and in the validation cohort (N=30, 9 responders patients (RCB=0) and 21 non responders (RCB > 0)) was performed to validate candidate genes obtained by Ilumina assay. Pyrosequencing results were calculated for each gene taking into account the methylation values from CpG identified in the Illumina assay and others from CpGs located close to them in order to obtain a more consistent result. In the DC we replicated methylation data in LOC641519/LEF1 gene (p=0.02) and HOXA5 gene (p=0.000l), where we observed a significant methylation higher in non-responders patients vs. responders in the same way that in Illumina study, and in FERD3L gene (p=0.04), TRIP10 (p=0.003) and CHL1 (p=0.03) where, as in Illumina assay, methylation was also significant higher in responders patients vs. non-responders. However, in genes EVC2 (p=0.07), CDKL2 (p=0.05) and TLX3 (p=0.07) replication was not significant but showed a trend toward a higher level of methylation in non-responders patients. Pyrosequencing in the VC validated the results for FERD3L gene (p= 0.0087) with high methylation in responders group vs. non-responders. For the final analysis performed in the whole cohort (DC+VC) all the genes with a p value <0.20 were selected. Apart from FERD3L the other gene under this threshold was TRIP 10 with a p=0.19 (see Figure 2).

Example 17. FERD3L methylation and gene expression in TNBC cell lines.

FERD3L methylation was studied by pyrosequencing in a set of TNBC cell lines and then gene expression studies by qPCR were performed. The aim was to correlate FERD3L methylation levels in each cell line with its gene expression level and corroborate results obtained in TNBC patients. It was observed that FERD3L gene was methylated in all the cell lines studied with methylation levels always higher than 40%. FERD3L expression correlated with the methylation detected as was expected, showing a low gene expression when methylation was high. Thereby, MDA-MB-231 cell line showed the lowest methylation level for FERD3L and correlated with the highest level of gene expression. Conversely, HCC1143 cell line that showed the higher methylation was the one with the lowest gene expression level (Figure 3A).

Example 18. FERD3L demethylation assays and gene expression in TNBC cell lines. As we can observe in Figure 3 A, FERD3L gene was methylated in MDA-MB- 436 cell line (~60%). In order to modify FERD3L methylation status and check if this change also affect to gene expression level, an assay with the MDA-MB-436 cell line treated with AZA demethylation agent (5-aza-2'-deoxycytidine) was performed. It was observed that AZA treatment modified FERD3L methylation in MDA-MB- 436 cell line inducing a decrease when compared it with control cells (p=0.05). As expected, this change in FERD3L methylation was correlated with a significant increase (p= 0.0022) in FERD3L expression (Figure 3B). High FERD3L expression levels correlates with low gene methylation in TNBC patients. It is well known that methylation can associated with changes in gene expression. Thus, when genes are methylated, for example genes FERD3L or TRIP 10, they usually show low gene expression levels. FERD3L expression in TNBC patients was studied to corroborate if there is a correlation between methylation and gene expression levels. When qPCR results obtained in the overall 54 patients studied was analyzed, it was observed a significant difference (p=0.04) in FERD3L gene expression with high expression in non-responders patients vs. responders (Figure 3C). Therefore, it suggests a correlation between low methylation and high gene expression in non-responders patients as we expected. To strengthen our data, we performed an analysis using The Cancer Genome Atlas dataset (TCGA) for breast cancer. In a subset of 713 patients we correlated FERD3L methylation with FERD3L expression and, according with our data, we found low gene expression when FERD3L was methylated. It was detected both when all CpGs in the FERD3L gene promoter were included and when only analyzed the CpG eg 10043037 validated for FERD3L gene in the study (Figure 3D).

Example 19. Statistical model to predict response to neoadyuvant treatment in TNBC patients.

Once the use of the methylation status of the genes FERD3L and TRIP 10 was validated for predicting the response to NAC (RCB = 0) or non-response to NAC (RCB > 0), a statistical model based on the Akaike Information criterion (AIC) was designed and ROC curve was construed. Such as it can be seen in Figure 4, a ROC curve with AUC=0.9056 (095% 0.805- 1) was finally obtained with a model wherein the methylation status of both genes FERD3L and TRIP 10 was measured.

Claims

1. In vitro method for the prediction of response to neoadjuvant chemotherapy in triple negative breast cancer patients, which comprises determining the methylation status or the level of expression of at least the genes FERD3L and TRIP10 in a biological sample obtained from the patient, wherein a higher level of methylation or a lower expression level of at least the genes FERD3L and TRIP 10, as compared with a reference level of methylation or expression level of the genes FERD3L and TRIP10 measured in non-responder patients, is indicative of response to neoadjuvant chemotherapy in triple negative breast cancer patients.

2. In vitro method for selecting a therapy for triple negative breast cancer patients which comprises determining the methylation status or the level of expression of at least the genes FERD3L and TRIP10 in a biological sample obtained from the patient, wherein a higher level of methylation or a lower expression level of the genes FERD3L and TRIP 10, as compared with a reference level of methylation or the expression level of the genes FERD3L and TRIP 10 measured in non-responder patients, is indicative of response to neoadjuvant chemotherapy in triple negative breast cancer patients.

3. In vitro method, according to any of the claims 1 or 2, wherein the neoadjuvant chemotherapy is non-platinum neoadjuvant chemotherapy comprising taxanes and/or anthracy clines; or platinum-based chemotherapy.

4. In vitro method, according to any of the claims 1 to 3, which comprises determining the methylation status of a CpG in the promoter region of the gene.

5. In vitro method, according to any of the claims 1 to 4, which comprises determining the methylation status of the promoter region CpG cgl0043037 of the gene FERD3L and/or the methylation status of the promoter region CpG cg02085507 of the gene TRIP 10.

6. In vitro method, according to any of the claims 1 to 5, which comprises determining the methylation status of a promoter region CpG which is up to 100 base pairs upstream or downstream from the CpG cgl0043037 of the gene FERD3L and/or from the CpG cg02085507 of the gene TRIP10, preferably between 1 and 80 nucleotides, and more preferably between 1 and 60 nucleotides.

7. In vitro use of the methylation status of the gene FERD3L and TRIP10 for the prediction of response to neoadjuvant chemotherapy in triple negative breast cancer patients.

8. In vitro use of the methylation status of the gene FERD3L and TRIP 10 for selecting neoadjuvant chemotherapy as a treatment for triple negative breast cancer patients.

9. In vitro use, according to any of the claims 7 or 8, wherein the neoadjuvant chemotherapy is non-platinum neoadjuvant chemotherapy comprising taxanes and/or anthracy clines; or platinum-based chemotherapy.

10. In vitro use, according to any of the claims 7 to 9, which comprises determining the methylation status of at least a CpG in the promoter region of the gene.

11. In vitro use, according to any of the claims 7 to 10, which comprises determining the methylation status of the promoter region CpG cgl0043037 of the gene FERD3L and/or the methylation status of the promoter region CpG cg02085507 of the gene TRIP 10.

12.7/7 vitro use, according to any of the claims 7 to 11, which comprises determining the methylation status of a promoter region CpG which is up to 100 base pairs upstream or downstream from the CpG cgl0043037 of the gene FERD3L and/or from the CpG cg02085507 of the gene TRIP10, preferably between 1 and 80 nucleotides, and more preferably between 1 and 60 nucleotides.

13. Kit comprising reagents for the determination of the methylation status of at least the genes FERD3L and TRIP 10.

14. Use of the kit, according to claim 13, for the prediction of response to neoadjuvant chemotherapy in triple negative breast cancer patients.

15. Use of the kit, according to claim 13, for selecting neoadjuvant chemotherapy as a treatment for triple negative breast cancer patients.