WO2023099890A1 - Method of prognosis - Google Patents

Abstract

The present invention relates to a method of predicting resistance to chemotherapy in a subject having triple negative breast cancer (TNBC), wherein said method comprises: (a) providing a biological sample from said subject; and (b) determining the expression levels of each member of a gene panel, said gene panel comprising at least 10 genes selected from the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1, DSC3, and AMIGO2 in said biological sample; wherein the determined expression levels of said genes of said gene panel is used to determine the likelihood of resistance to said chemotherapy, wherein said gene panel comprises one, two, three, four, five or six of ITGB1, RBFOX2, DST, RCAN1, c9orf3 and ACTA2.

Description

METHOD OF PROGNOSIS

Field Of The Invention

The present invention relates to methods of characterizing breast cancer. In particular, the invention relates to methods of predicting resistance to chemotherapy in subjects with triple negative breast cancer.

Background to the Invention

Triple-negative breast cancer (TNBC) lacks expression of estrogen and progesterone receptors and HER2 protein and accounts for ~20% of all breast cancer cases. TNBC is highly aggressive, and is associated with a poor prognosis with 40% mortality within the first 5 years after diagnosis ¹’³. Chemotherapy is the first line of treatment option for patients with TNBC in neoadjuvant, adjuvant, or metastatic settings. While chemotherapy is effective in some TNBC patients, nearly half of the patients develop resistance to chemotherapy, which results in poor overall survival ⁴⁵. Moreover, unlike in ER-positive cancers, eliminating the majority of the bulk population of cancer cells has relatively little impact on the clinical outcomes for TNBC ⁴⁵. This suggests that the minor, escaping cell subpopulations may underlie TNBC aggressiveness including chemoresistance.

There is increasing evidence that epithelial to mesenchymal transition (EMT) contributes to TNBC aggressiveness by conferring metastatic ability and resistance to therapies ^6-11. Indeed, the majority of breast cancer deaths (90%) are caused by tumor invasion and metastasis, which are the two key features related to EMT ¹². Multiple lines of evidence suggest that intratumoral heterogeneity drives these aspects of TNBC aggressiveness ^13-18. Despite advances, it is not known whether transcriptionally distinct EMT subpopulations contribute to therapy resistance in TNBCs. Furthermore, whether the gene expression program defining these subpopulations can be developed into biomarkers has not been explored. Currently available signature gene panels for predicting chemotherapy response such as Oncotype DX (21 genes, RT-PCR) ^19-21 , Endopredict (12 genes, RT-PCR) ²² , PROSIGNA (50 genes, RT-PCR) ²³ are designed for ER-positive breast cancers and are not effective in the prediction of resistance to chemotherapy in the treatment of TNBC. Increasing evidence suggests that there may exist similar gene signatures for TNBC ^24-3°. For example, while one study showed a relationship of proliferation and immune-related genes, another one showed strom a- related genes to be predictive of Neoadjuvant chemotherapy (NAC) response in TNBC patients. However, previous efforts to identify chemoresistance genes have failed in capturing those expressed in small subpopulations associated with chemoresistance in TNBC. None of these signatures reach a high prediction accuracy, limiting their clinical utility. Unfortunately, no targeted therapies are available in the clinical settings for the treatment of TNBCs, except PARP inhibitors in germline BRCA1/2-mutated tumours. Chemotherapy is the only treatment option for most TNBC patients. There is a highly unmet need to characterize subpopulations that drive therapy resistance in TNBC and facilitate the development of biomarkers to predict therapy response to enable precision medicine and open new avenues for targeted therapeutic intervention for this deadly cancer.

Summary Of The Invention

In the present study, the inventors performed an integrated analysis of gene expression profiles derived from scRNA-seq, bulk RNA-seq, and Microarray data from TNBC tumors treated with chemotherapy and identified subpopulations of cells that associate with key aspects of TNBC aggressiveness such as chemoresistance. The inventors show that the identified signature genes can accurately predict response to NAC in primary as well as advanced stage TNBC (lymph-node positive), with the prediction accuracy greatly improved over that of known gene expression signatures for TNBC.

Accordingly in a first aspect, the present invention provides a method of predicting resistance to chemotherapy in a subject having triple negative breast cancer (TNBC), wherein said method comprises: a) providing a biological sample from said subject; and b) determining the expression levels of each member of a gene panel in said biological sample, said gene panel comprising at least 10 genes selected from the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3/AOPEP, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1, DSC3, and AMIGO2; wherein the determined expression levels of said at genes of said gene panel is used to determine the likelihood of resistance to said chemotherapy.

In a particular embodiment of the invention, the gene panel comprises the following 20 genes: ITGB1 , RBFOX2, DST, RCAN1, c9orf3/AOPEP, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1, STOM, COL6A1, DSC3, and AMIGO2.

In a particular embodiment of the invention, the gene panel consists of the following 20 genes: ITGB1 , RBFOX2, DST, RCAN1, c9orf3/AOPEP, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1, STOM, COL6A1, DSC3, and AMIGO2.

As described in the examples, the inventors have shown that such a gene panel (gene signature) can predict response to chemotherapy in TNBC significantly more accurately than known gene expression signatures for TNBC. Moreover, the inventors have further shown that smaller gene panels comprising fewer of the 20 genes still outperform known gene signatures. For example, as shown in the Examples, even when the five genes having greatest effect on the performance of the gene panel are not included, the overall performance of the gene signature is still greater than that of known gene signatures. Moreover, given the demonstration that when the removal of the five genes having the least effect on the performance of the gene panel does not significantly reduce the performance of the gene panel, the invention extends to gene panels which do not include all 20 of ITGB1 , RBFOX2, DST, RCAN1, c9orf3/AOPEP, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1, DSC3, and AMIGO2. Accordingly, in another embodiment of the invention, said gene panel comprises at least 10 genes selected from the group consisting of ITGB1, RBFOX2, DST, RCAN1 , c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, and ACTN1, wherein said gene panel comprises one, two, three, four, five or six of ITGB1, RBFOX2, DST, RCAN1 , c9orf3 and ACTA2.

In another embodiment, the gene panel comprises at least 15 genes selected from the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1, DSC3, and AMIGO2, wherein said gene panel comprises one, two, three, four, five or six of ITGB1, RBFOX2, DST, RCAN1 , c9orf3 and ACTA2

In another embodiment, the gene panel comprises at least the fifteen genes of the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, and ACTN1.

In such an embodiment, said gene panel may comprise (i) at least the fifteen genes of the group consisting of ITGB1, RBFOX2, DST, RCAN1 , c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, and ACTNI ; and optionally (ii) one or more of the genes of the group consisting of SFRP1 , STOM, COL6A1 , DSC3, and AMIGO2.The likelihood of resistance to chemotherapy may be determined by comparing the expression levels of said genes of the gene panel relative to reference amounts of said genes. Such control cells may chemosensitive TNBC cells, for example chemosensitive TNBC cells from patients shown to have pathological complete response (pCR) to said chemotherapy.

In a preferred embodiment of the invention, each gene of the gene panel has significantly different level of expression in TNBC cells from patients with residual disease vs TNBC cells from patients with pCR (e.g. the difference having a p value of less than 0.05). Some genes of the gene panel may have significantly higher expression in TNBC cells from patients with residual disease compared to the expression of these genes in TNBC cells from patients with pCR. Other genes may have significantly lower expression in TNBC cells from patients with residual disease compared to the expression of these genes in TNBC cells from patients with pCR.

In embodiments of the invention, the likelihood of resistance to said chemotherapy may be determined by applying the expression levels to a predictive model which relates expression levels of said genes of the gene panel with resistance to chemotherapy against triple negative breast cancer.

Applying the expression levels to such a predictive model may comprise weighting the expression levels of said genes of the gene panel according to a predetermined ranking of said genes of the gene panel. In one embodiment, the weighting of the expression levels of the genes may be determined by the coefficient value derived for each gene using LASSO regression.

In one such embodiment, where the gene panel comprises all 20 genes of the gene panel of the first aspect of the invention, the genes may be ranked in order of greatest to least predictive power, for example as shown in Figure 4C.

Detailed Description

Unless otherwise defined, all technical and scientific terms used herein have the meaning commonly understood by a person who is skilled in the art in the field of the present invention.

Throughout the specification, unless the context demands otherwise, the terms “comprise” or “include”, or variations such as “comprises” or “comprising”, “includes” or “including” will be understood to imply the inclusion of a stated integer or group of integers, but not the exclusion of any other integer or group of integers.

As used herein, terms such as "a", "an" and "the" include singular and plural referents unless the context clearly demands otherwise. Thus, for example, reference to "an active agent" or "a pharmacologically active agent" includes a single active agent as well as two or more different active agents in combination, while references to "a carrier" includes mixtures of two or more carriers as well as a single carrier, and the like.

The present invention is based on the identification of a specific gene signature which the inventors have shown can be used to predict with high accuracy 5 resistance against chemotherapy in TNBC patients. Genes which may be used in the gene signatures of the invention may comprise the following 20 genes: ITGB1, RBFOX2, DST, RCAN1, c9orf3/AOPEP, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1 , DSC3, and AMIGO2. Examples of nucleic acid sequences for these genes are shown in the sequence 10 listing with sequences for each of the genes as follows: ITGB1 (SEQ ID NO: 1) RBFOX2 (SEQ ID NO: 4), DST (SEQ ID NO: 5), RCAN1 (SEQ ID NO: 10), c9orf3/AOPEP (SEQ ID NO: 17), ACTA2 (SEQ ID NO: 9), S100B (SEQ ID NO: 8), LY6E (SEQ ID NO: 19), CTNNAL1 (SEQ ID NO: 18), PRNP (SEQ ID NO: 2), TIMP3 (SEQ ID NO: 16), CD63 (SEQ ID NO: 3), IFI16 (SEQ ID NO: 12), NFIB (SEQ ID NO: 15 13), ACTN1 (SEQ ID NO: 6), SFRP1 (SEQ ID NO: 11), STOM (SEQ ID NO: 15),

COL6A1 (SEQ ID NO: 14), DSC3 (SEQ ID NO: 20), and AMIGO2 (SEQ ID NO: 7).

Genes which may be used in the gene signatures are listed in Figure 7. Examples of suitable Affymetric probe sequences which may be used to identify target sequences 20 of the genes are listed in Table 1.

Table 1 :

It should be understood that for any particular gene of the gene signature, a number of target sequences may be used. The skilled person will readily be able to identify suitable target sequences for each gene and likewise will readily be able to design suitable probes/primers based on the sequences of the genes and/or of individual

5 target sequences. Accordingly, the recited probes and target sequences recited for each gene of the gene panel should be considered as non-limiting examples.

In some embodiments of the invention, the gene panel may comprise c9orf3. In some embodiments of the invention, the gene panel may comprise CD63. In some embodiments of the invention, the gene panel may comprise STOM.

10 In one embodiment, the gene panel comprises c9orf3 and at least 9, for example, 10, 11, 12, 13, 14, 15, 16, 17 , 18, or all 19 of the genes selected from the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1, STOM, COL6A1, DSC3, and AMIGO2. In one embodiment, the gene panel comprises CD63 and at least 9, for example, 10, 11, 12, 13, 14, 15, 16, 17 , 18, or all 19 of the genes selected from the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1, DSC3, and AMIGO2.

In one embodiment, the gene panel comprises STOM and at least 9, for example, 10, 11, 12, 13, 14, 15, 16, 17 , 18, or all 19 of the genes selected from the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1 , COL6A1 , DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel comprises at least two or three genes selected from the group consisting of: c9orf3, CD63, and STOM. Thus in one embodiment, the gene panel may comprise c9orf3 and CD63. In another embodiment, the gene panel may comprise c9orf3 and STOM. In another embodiment, the gene panel may comprise CD63 and STOM. In one embodiment of the invention, the gene panel comprises c9orf3, CD63, and STOM.

In one embodiment, the gene panel comprises two or three of the genes selected from c9orf3, CD63 and STOM and at least 8, for example, 9, 10, 11 , 12, 13, 14, 15, 16, or 17 of the genes selected from the group consisting of ITGB1, RBFOX2, DST, RCAN1 , ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, IFI16, NFIB, ACTN1, SFRP1, COL6A1, DSC3, AMIGO2, KRT17, ACTG2, MYLK, ANXA1, CNN3, CAV2, MSRB3, and TNG

In some embodiments of the invention, the gene panel may comprise one, two, three, four, five or six of ITGB1, RBFOX2, DST, RCAN1 , c9orf3 and ACTA2. In some such embodiments, the gene panel may comprise four, five or six of ITGB1, RBFOX2, DST, RCAN1 , c9orf3 and ACTA2 together with one, two or three genes selected from the group consisting of: c9orf3, CD63, and STOM. In embodiments of the invention, the gene panel may comprise, or consist of, at least 12, 13, 14, 15, 16, 17, 18, or 19 genes of the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1 , DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel comprises (i) at least the twelve genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, and CD63; and optionally (ii) one or more of the genes of the group consisting of IFI16, NFIB, ACTN1, SFRP1 , STOM, COL6A1 , DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel comprises (i) at least the thirteen genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, and IFI16; and optionally (ii) one or more of the genes of the group consisting of NFIB, ACTN1, SFRP1, STOM, COL6A1 , DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel comprises (i) at least the fourteen genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1 , c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16,and NFIB; and optionally (ii) one or more of the genes of the group consisting of ACTN 1 , SFRP1 , STOM, COL6A1 , DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel comprises (i) at least the fifteen genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, and ACTN1 ; and optionally (ii) one or more of the genes of the group consisting of SFRP1 , STOM, COL6A1 , DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel comprises (i) at least the sixteen genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1,and SFRP1; and optionally (ii) one or more of the genes of the group consisting of STOM, COL6A1 , DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel comprises (i) at least the seventeen genes of the group consisting of ITGB1, RBFOX2, DST, RCAN1 , c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, and STOM; and optionally (ii) one or more of the genes of the group consisting of COL6A1 , DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel comprises (i) at least the eighteen genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1 , c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1 , STOM, and COL6A1 ; and optionally (ii) one or more of the genes of the group consisting of DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel comprises (i) at least the nineteen genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1 , c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1 , STOM, COL6A1, and DSC3; and optionally (ii) AMIGO2.

In a particular embodiment of the present invention, the gene panel comprises the twenty genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1 , STOM, COL6A1, DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel is limited to genes selected from the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1 , DSC3, and AMIGO2, i.e. , the genes of which expression is determined are limited to only ten or more of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1, STOM, COL6A1 , DSC3, and AMIGO2. In one embodiment of the invention, the gene panel consists of the twelve genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, and CD63.

In one embodiment of the invention, the gene panel consists of the thirteen genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, and IFI16.

In one embodiment of the invention, the gene panel consists of the fourteen genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, and NFIB.

In one embodiment of the invention, the gene panel consists of the fifteen genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, NFIB, and ACTN1.

In one embodiment of the invention, the gene panel consists of the sixteen genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, and SFRP1.

In one embodiment of the invention, the gene panel consists of the seventeen genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1 , and STOM.

In one embodiment of the invention, the gene panel consists of the eighteen genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1 , STOM, and COL6A1.

In one embodiment of the invention, the gene panel consists of the nineteen genes of the group consisting of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1, and DSC3. In one embodiment of the invention, the gene panel consists of the following genes: ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1, DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel consists of the following genes: RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1 , STOM, COL6A1 , DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel consists of the following genes: DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1, STOM, COL6A1, DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel consists of the following genes: RCAN1 , c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1 , DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel consists of the following genes: c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1 , STOM, COL6A1, DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel consists of the following genes: ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1 , STOM, COL6A1, DSC3, and AMIGO2.

In one embodiment of the invention, the gene panel consists of the following genes: ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1 , STOM, COL6A1, and DSC3,.

In one embodiment of the invention, the gene panel consists of the following genes: ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1 , STOM, and COL6A1. In one embodiment of the invention, the gene panel consists of the following genes: ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1, and STOM.

In one embodiment of the invention, the gene panel consists of the following genes: ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , and SFRP1.

In alternative embodiments, the gene panel may comprise, in addition to genes selected from the group ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1, DSC3, and AMIGO2, other biomarkers associated with TNBC. For example, in such embodiments the gene panel may additionally comprise one or more of the genes selected from KRT17, ACTG2, MYLK, ANXA1 , CNN3, CAV2, MSRB3, and TNC. In one embodiment of the invention, the gene panel is limited to genes selected from the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1, DSC3, AMIGO2, KRT17, ACTG2, MYLK, ANXA1 , CNN3, CAV2, MSRB3, and TNG i.e., the genes of which expression is determined are limited to only ten or more of ITGB1 , RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1, STOM, COL6A1, DSC3, and AMIGO2 and optionally one or more of the genes selected from KRT17, ACTG2, MYLK, ANXA1, CNN3, CAV2, MSRB3, and TNG.

The inventors have also identified the following genes as being particularly powerful in the TNBC gene signatures of the invention: MSH3, TSHZ2, SLC11A2, CTDSPL, and NDUFA6.

Accordingly, in embodiments of the invention, the gene panel may comprise one, two, three, four, or five of MSH3, TSHZ2, SLC11A2, CTDSPL, and NDUFA6. Examples of nucleic acid sequences for these genes are shown in the sequence listing as SEQ ID NOS 21, 22, 23, 24 and 25 respectively. In one embodiment, the gene panel comprises no more than 20 genes. In a particular embodiment of the invention, the gene panel consists of the following 20 genes: ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1, DSC3, and AMIGO2.

Determining Expression Levels

In the methods of the invention, the expression level of genes of the gene panel may be identified using any suitable method. Expression of the genes may be measured directly by measurement of mRNA expression via a microarray, Northern blotting, in situ RNA detection, RNA sequencing, PCR methods or any other suitable nucleic acid amplification technique. The methods may involve any suitable primers and/or probes, the design of which is within the routine knowledge of the skilled person. Primer design tools, such as the NCBI Primer-BLAST tool may be used. Typically, primers and/or probes will be approximately 15-25 nucleotides in length. Other methods used to determine gene expression include next generation sequencing methods employing, for example, pyrosequencing, Ion Torrent semiconductor sequencing, Illumina sequencing methods, single-molecule real-time sequencing or DNA nanoball sequencing. Alternatively, expression of the genes in the gene samples may be determined by in situ RNA detection using, for example, in situ hybridisation techniques to localise specific RNA sequences in a cell or section of tissue.

In addition to methods of determining expression of the genes of the gene panel which rely on measurement of mRNA directly, the expression levels of one or more of the genes of the gene panel may be determined by measurement of protein products of the genes in cells of a tissue sample. Such methods of determining expression of the genes at the protein level may include commonly known techniques such as Western blot, immunocytochemistry, immunoprecipitation, mass spectrometry, ELISA, etc. In one option, immunohistochemistry may be used to determine the level of proteins from particular genes of the gene panel in cells of a tissue sample of interest. In such methods, antibodies or aptamers directed to the protein of interest may be used to determine its level. Methods for generating such antibodies or aptamers are well known to the person skilled in the art. The antibody or aptamers used in such methods may be conjugated to a label, the label forming part of a detection agent. Such methods are well known in the art.

It is to be understood that, while in many cases it may be preferable for ease of operation to determine the expression level of each of the genes in a gene panel using the same technique, it is within the scope of the invention to use a number of different methods to determine expression levels of genes of the gene panel.

Reference Levels

In the method of the invention, the expression level of each member of the gene panel may be determined with reference to a reference or control from e.g. chemosensitive TNBC cell(s). A gene may be considered to be differentially expressed in a sample from a subject as compared to a control reference value e.g. from a non-tumour sample from a subject or group of subjects where the gene is expressed at a level which is significantly increased or significantly decreased compared to the control cell/reference value.

A reference level of a particular gene may be an absolute or relative amount or concentration of the gene, a presence or absence of the gene product, a range or amount of concentration of the gene product, a mean amount of the gene or gene product, and/or a median amount of or concentration of the gene or gene product. A “reference level” can also be a “standard curve reference level” based on a level of one or more of the genes determined from a population and plotted on appropriate axes to produce a reference curve. The reference curve may be tailored to particular populations of subjects, with reference levels varying with, for example, age group. A standard curve reference level may be determined from a group of reference levels from a group of subjects having a particular disease using statistical analysis, such as univariate or multivariate regression analysis, logistic regression analysis, linear regression analysis, etc. of the levels of such genes/biomarkers in samples from the group. Such reference levels may be adjusted to specific techniques used to measure levels of gene expression, where the gene expression levels may differ based on the specific technique that is used. In one embodiment of the invention, where the expression level of 10 or more genes of the gene panel is significantly different relative to the predetermined reference level for these genes, the sample may be defined as positive for the gene signature, i.e. the demonstration that at least 10 genes of the gene panel have significantly different levels of expression relative to reference levels in control cells for the corresponding genes is indicative that the sample is from TNBC cells which are resistant to chemotherapy. In another embodiment, the presence of 11 or more, 12 or more, 13 or more, 14 or more, 15 or more, 16 or more, 17 or more, 18 or more, or 19 or more of the genes of the gene panel being expressed at an significantly different level compared to the reference levels in control cells is indicative of the sample being a sample of TNBC cells which are chemotherapy resistant.

As noted above, some genes of the gene panel may have significantly higher expression in TNBC cells from patients with residual disease compared to the expression of these genes in TNBC cells from patients with pCR. Other genes may have significantly lower expression in TNBC cells from patients with residual disease compared to the expression of these genes in TNBC cells from patients with pCR. In embodiments of the present invention, where the gene panel comprises one or more, for example four, or more of the genes of the group consisting of RBFOX2, DST, RCAN1, c9orf3, ACTA2, LY6E, PRNP, TIMP3, CD63, ACTN1, STOM, COL6A1, DSC3, and AMIGO2, expression of said one or more of said genes at a significantly higher level than in control cells is indicative of the sample being a sample of TNBC cells which are chemotherapy resistant. In embodiments of the present invention, where the gene panel comprises one or more of the genes of the group consisting of ITGB1 , S100B, CTNNAL1, IFI16, NFIB, and SFRP1 , expression of said one or more of said genes at a significantly lower level than in control cells is indicative of the sample being a sample of TNBC cells which are chemotherapy resistant.

In a simple version of the method of the invention, the relative expression of each of the genes of the gene panel compared to reference expression levels of these genes from a control is combined to produce a compound gene signature score Thus, it is possible that, while some genes of the gene panel do not demonstrate significantly different levels of expression relative to the corresponding genes in the control(s), sufficient numbers of the genes of the gene panel demonstrate significantly different levels of expression such that the compound gene signature score reaches a threshold level which is indicative of resistance to chemotherapy.

As described in the examples, each of the genes which may be used in the gene panels in the methods of the invention have been assessed for predictive power in relation to the overall gene signature, with the genes ranked in order of influence on the gene signature performance (see Figure 4C). Each of the genes of the 20 gene panel comprising ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1, STOM, COL6A1, DSC3, and AMIGO2 has thus been assigned a Predictive Score indicative of its influence on the performance of the 20 gene signature, with a higher value indicative of greater reduction in performance of the gene signature upon the exclusion of that gene from the gene signature. The ranking of each gene in the gene panel may be taken into account in determining the overall compound expression score for the gene panel to determine the likelihood of resistance to chemotherapy in TNBC cells.

Accordingly, in one embodiment of the invention, in the determining of the likelihood of resistance to chemotherapy against triple negative breast cancer, the compound expression score is calculated employing a weighting of the expression levels of the genes of the gene panel according to the ranking of the genes of the gene panel. The ranking may be the rank position, in which a ranking of “1” is indicative that the gene has the most negative impact on performance when removed with rankings of higher numbers indicating less impact on performance when removed from the gene signature. Thus, the genes of a gene signature of 20 genes may be ranked from 1 to 20, with the gene having the greatest positive effect on the predictive power of the gene signature having rank position 1 , the gene with the second highest positive effect on the predictive power of the gene signature having rank position 2, etc. Alternatively, any formula used to determine the compound expression score may take into account the ranking score, such as the ranking scores as listed for the 20 gene signature in Figure 4C. In such case, the higher the ranking score of a gene, the greater the impact on performance on the gene signature when that gene is removed from the gene signature The ranking score may change depending on the total number of genes in a gene panel. The combined expression level, as appropriately weighted according to ranking score, may thus be used to determine the final gene signature score.

The skilled person will appreciate that the genes included in the gene signature may thus carry unequal weight in the determination of the likelihood of resistance to chemotherapy in TNBC cells. Accordingly, the weighting/rank of each gene in the score may be taken into account in the calculation of the overall gene signature score.

In one embodiment, the gene signature score may be determined using any suitable weighting formula. In one embodiment, a suitable weighting formula may make use of regression analysis. Any suitable predictive package may be used. For example one such predictive package which may be used is the caret R package. In summary it takes the absolute value of the final coefficients and ranks them based on their importance in the model. The main function is varlmp() which is described as a generic method for calculating variable importance for objects produced by “train” and method specific methods (https77www.rdocumentation.Org/packages/caret/versions/6.0-90/topics/varlmp) . It can be applied to multiple models including linear, random forrest and glmnet where it calculates the importance of the outputs and ranks them. In the case of glmnet the outputs are coefficients. Thus, in one embodiment, the varlmp function in the caret package may be used to rank the importance of genes of the gene signature.

In one such embodiment, the following code may used for gene ranking varlmp <- function (object, lambda = NULL, ...) { beta <- predict(object, s = lambda, type = "coef") if(is.list(beta)) { out <- do.call("cbind", lapply(beta, function(x) x[, 1])) out <- as.data.frame(out, stringsAsFactors = TRUE)

} else out <- data.frame(Overall = beta[, 1]) out <- abs(out[rownames(out) != "(Intercept)",, drop = FALSE]) out

} varlmp(cv.lassoModel, lambda = cv.lassoModel$lambda.min)

The compound gene signature score may be compared to a reference value which may be derived from, for example, a training set of patent data in which, for example, the threshold is established to indicate a total score which is indicative of whether or not the cells of the sample are chemotherapy resistant.

The performance of a particular gene signature may be determined in any suitable way. In one embodiment, the performance is assessed using Area Under the Curve (AUC). AUC refers to the area under the curve of a Receiver Operating Characteristic (ROC) curve. A higher AUC value is indicative of greater capacity of the gene signature (gene panel) to predict classes correctly, in this case or not a sample falls in one or another of two groups of interest, e.g. in the present case, the ability to predict whether or not cells of a sample are chemotherapy resistant.

AUC is typically measured as a value between 0 and 1 where an AUC of 0 is indicative that the model’s predictions are 100% wrong and an AUC of 1 is indicative that the model’s predictions are 100% correct. AUC may also be presented on a scale of 0% to 100%, with an AUC of 0% is indicative that the model’s predictions are 100% wrong and an AUC of 100% is indicative that the model’s predictions are 100%

In embodiments of the invention, the AUC associated with a gene panel of the invention is greater than 0.800. Optionally, the AUC associated with a gene panel of the invention is greater than 0.830, for example greater than 0.840, such as greater than 0.850, greater than 0.860, greater than 0.870, such as greater than 0.880, greater than 0.890 or greater than 0.900.

Other methods may be used to characterise whether or not a sample falls within the chemoresistant TNBC category other than by determining AUC and use of Logistic LASSO regression. For example, other methods which could be used include classification and regression trees, Random Forests, Multivariate Adaptive Regression Splines, Support vector machines and Decision Tree etc.

Chemotherapy

The methods of the invention enable the determination as to whether a patient’s TNBC is likely to be resistant to a chemotherapy treatment. In one embodiment, the chemotherapy comprises one, two, three or more of the group consisting of anthracyclines (for example Adriamycin (doxorubicin)), alkylating agents (for example Cyclophosphamide), taxanes (for example taxol (Docetaxel)), or antimetabolites (such as fluorouracil (5-Fll).

In one embodiment, the chemotherapy treatment is a combined chemotherapy treatment regimen which comprises or consists of an anthracycline and an alkylating agent. In one such embodiment, the combined chemotherapy treatment regimen consists of Adriamycin (doxorubicin) and cyclophosphamide. In another embodiment, the combined chemotherapy treatment regimen consists of Adriamycin (doxorubicin), cyclophosphamide, and paclitaxel.

In another embodiment, the chemotherapy treatment is a combined chemotherapy treatment regimen which comprises or consists of an anthracycline, an alkylating agent, and an antimetabolite. In one such embodiment, the combined chemotherapy treatment regimen comprises or consists of 5-Fll, Adriamycin, and Cyclophosphamide. In another embodiment, the combined chemotherapy treatment regimen consists of paclitaxel, 5-Fll, Adriamycin, and Cyclophosphamide,

Brief Description of the Figures

Embodiments of the present invention will now be described, by way of example only, with reference to the accompanying figures:

Figure 1 : Single-cell transcriptomic analysis reveals cell populations associated with TNBC aggressiveness. A) Schematic workflow of the study. The inventors utilised scRNA-seq datasets of treatment naive primary and chemotherapy-treated TNBC patients and identified small cell populations associated with chemoresistance-like characteristics. Further, genes defining these subpopulations were validated in bulk RNA-seq of tumors chemoresistance tumors and utilised for building a predictive model in stratifying patients with residual disease and pathological complete response. B) scRNA-seq data analysis of six TNBC tumors from dataset 1 shows cellular heterogeneity profile. The genes defining each cluster were annotated against cellmarker database to assign cell types identity to each cluster. C) Dot plot showing top 5 ranked markers defining each cluster. D) Violin plot showing expression of signature genes associated with metastasis (top plot) signatures. Metastasis signature of 49 genes from Lawson et al, 2015 was utilised and their expression was plot across each cluster, Chemoresistance signature (bottom plot) of 143 genes was used from Balko et al, 2012 and their expression was plotted across each cell type. Statistical significance of expression levels were calculated using paired t-test between the clusters. E) Venn diagram showing EMT cell populations defining genes, reproducible across three datasets. EMT cell type defining genes overlapped between all three datasets and genes evident in at least two datasets was considered as reproducible signature genes. F) Kaplan Meier survival analysis plot showing correlation of signature gene expression with 5-year relapse-free survival (RFS) in TNBC patients. Analysis was performed using mean expression levels of all 101 signature genes in 417 TNBC patients using km pl otter.

Figure 2: Distinct EMT-related pathways pre-exist to confer chemoresistance in TNBC.

Expression analysis of signature genes in 7 TNBC tumors (collected pre-treatment and during the surgical excision after six cycles of NAC (docetaxel and epirubicin) (post-treatment) were analysed. A) LIMAP plot of chemo-sensitive and chemoresistance patients showing indicating pre and post-treated cell clusters. B) LIMAP feature plot (left panel) represents average expression of signature genes across chemo-sensitive and chemo-resistance clusters. The average expression profile of signature genes between the groups is indicated in the violin plot, (right panel). The expression level significance was calculated using paired t-test. C) Violin plot showing genes that are highly expressed in chemoresistance groups and have no or minimal expression in the chemosensitive clusters. Genes were plotted on the scRNA-seq dataset of clonal extinct (chemosensitive) and persist (chemoresistance) patients. D) The dot plot shows the enrichment score (NES) of the top transcription factor binding motifs identified by iRegulon from the promoter regions of the EMT defining signatures genes of the aggressive subpopulations. The criteria set for motif enrichment analysis were as follows: identity between orthologous genes > 0.0, FDR on motif similarity <0.001, and TF motifs with normalized enrichment score (NES) > 3. The ranking option for Motif collection was set to 10 K (9,713 PWMs) and a putative regulatory region of 20 kb centered around TSS was selected for the analysis. E) The Heatmap shows enrichment of 50 hallmark signature pathways in chemosensitive and chemoresistance subpopulations using the Molecular Signatures Database (MSigDB). The pathways highlighted in the red and green boxes are enriched in chemoresistance and chemosensitive subpopulations. F) Boxplot showing average expression of signature genes in bulk RNA-seq datasets of 75 chemotherapy-treated TNBC patients.

Figure 3: Chemoresistance signature genes are enriched for EMT-related processes and tumors of mesenchymal features. A) Gene ontology analysis of reproducible signature genes enriched for biological processes associated with EMT. B) Boxplot of TCGA breast cancer cohort showing average expression of signature genes across different subtypes. Expression profile of signature genes extracted from TCGA breast cancer cohort and each tumor classified into four subtypes i.e. LuminalA, LuminalB, Her2, and TNBC based on ER, PR, Her2 status. Next, the average expression of these signature genes was plotted across subtypes of breast cancers. Statistical significance was calculated using paired Wilcoxon signed-rank test between the subtypes. C) Expression of signature genes in EMT-High and EMT- Low TNBC tumors. Correlation plot showing the classification of TCGA TNBC tumors into EMT-high and EMT-low groups based on the expression of 12 established EMT markers. Size and color represent correlation significance. Circle with blue and red colors showing positive and negative correlations between the epithelial and mesenchymal genes. D) Boxplot showing average expression of signature genes in EMT-High and EMT-Low TNBC (TCGA) cohort. E) The Heatmap showing expression dynamics of signature genes across different EMT timepoints (TGF-b treated) of mammary epithelial cells (HMLE) RNA-seq. The time point labeled with dO are untreated and d1-d20 are different EMT timepoints treated with TGF-b at day 1 to day20. A cluster of genes represented by C is on the right side of the heatmap representing EMT induction time-specific genes.

Figure 4: A 20-gene pane! can accurately predict chemotherapy response in TNBC patients. The predictive model development and validation with 20 gene expression levels using the least absolute shrinkage and selection operator (LASSO) regression method. A) The selection of tuning parameter (A) in the LASSO model based on the 10-fold cross-validation. The plots are showing a cross-validation curve (red dotted line) along with mean binomial deviance against a range of Log(A). The vertical dotted lines represent lambda, min and lambda.1se. This panel shows the changes of partial likelihood deviance with A values. The 20 genes were selected according to the most regularized model such that the error is within one standard error of the minimum. B) The coefficients from the Lasso fit represent the contributions of the 20 genes expression in the model and. C) Ranking of genes based on the lamda score obtained from glmnet. D) The receiver operating characteristic (ROC) curves of TNBC validation dataset for pCR and RD prediction using 20 gene expressions as a feature. E) The ROC curves were generated for all 20 gene features (red-top curve); after removal of one gene (cadetblue-middle curve); and removal of 5 genes (purple-lower curve) to assess the impact of the combination of genes on model performance. The ROC curve highlighted with the red line used all 20 genes whereas the model with reduced gene sets (highlighted with cadetblue and purple line) shown a difference in discriminating ability in predicting pCR vs RD in TNBC. F) The ROC curve showing the comparative performance in the TNBC validation dataset of the inventors’ model (QUB EMT panel -AUC 90.3%) compared to that of six published signatures. G) The receiver operating characteristic (ROC) curves of the lymph node-positive TNBC validation dataset to predict pCR and RD. H) Kaplan Meier survival analysis plot showing correlation of 20 gene expression with 5-year relapse-free survival (RFS) in TNBC (top panel) (n=417) and Lymph node-positive (bottom panel) (n=143) patients. Analysis was performed using mean expression levels of all 20 genes using kmplotter. I) Panel I shows ROC curves were generated for gene panels after removal of one to five of the top ranked genes (top panel); and after removal of one to five of the lower ranked genes (bottom panel). The gene(s) recited above each ROC curve is/are the gene(s) removed from the gene panel for the assessment; e.g, in the top panel, the first ROC curve is for a 19 gene gene panel consisting of the genes shown in the list on the left excluding ITGB1, the second ROC curve is for an 18 gene gene panel consisting of the genes shown in the list on the left excluding ITGB1 and RBFOX2, etc.

Figure 5. A) scRNA-seq data analysis of pre-treated TNBC patients identified a similar subpopulation in two independent TNBC datasets. The genes defining each cluster were annotated against the cellmarker database and cell-type identities were assigned to each cluster. B) Dot plot showing top 5 ranked genes defining each cluster of each TNBC scRNA-seq dataset. C) Expression of metastasis-associated genes in pre-treated TNBC patients scRNA-seq datasets. Metastasis signature of 49 genes was used by Lawson et al, 2015 and their average expression was plotted across each cluster in both datasets, D) Chemoresistance signature of 143 genes was used from Balko et al, 2012 and their average expression was plotted across each cell types in both datasets.

Figure 6. A) Violin plot showing average expression of signature genes across clusters of both TNBC datasets. The average expression of all 101 signature genes was plotted in all three primary TNBC scRNA-seq datasets and confirmed their activation in similar subpopulations of basal epithelial cells. B) LIMAP plot showing expression of established EMT marker SNAI2 expression across TNBC clusters confirming subpopulation underlying EMT. C) LIMAP plot showing total cluster identified in the single-cell datasets of 7 TNBC patients pre and post-chemotherapy. D) TF-targeted regulatory network showing top 5 transcription factors and their target genes involved in transcriptional regulation of aggressive subpopulations. The regulatory networks were generated by the iRegulon (version 1.3) package and the TF-gene network was plotted using Cytoscape (version: 3.8). The genes in the hexagonal box indicate transcription factors (node in the network), and their regulatory targets are shown in the oval shape. The line shows the connection of each TF (node) with their target genes (edges) in the network.

Figure 7 lists genes for use in the gene signatures of the invention.

EXAMPLES

METHODS

Identification of chemoresistance and metastasis cell populations using single-cell transcriptome profiling of TNBC tumors

The inventors’ first goal was to uncover cellular diversity in TNBC and assess the presence of cells exhibiting chemoresistance and metastasis-like features (Fig.lA). For this, the inventors analyzed the single-cell RNA-seq profile of more than 10000 cells (10,779 cells) derived from 33 TNBC patients, including patients treated with neoadjuvant chemotherapy (NAC) from four different studies ^14115131132. All four datasets were analyzed with the uniform parameters in the Seurat package (version 4.0.1), where cell pre-processing was done, and cells were removed having unique feature counts >2,500 or <200 and had mitochondrial reads >5%. Further downstream analysis was performed using the same Seurat package, in which data normalization was performed followed by variable feature selection, dimensionality reduction, and clustering. The clusters were identified with the LIMAP reduction method. For assigning cell type identity to each cluster, the inventors used manually curated marker lists of breast cell types from the CellMarker database ³³.

Identification of disease aggressiveness associated (metastasis and chemoresistant) subpopulations

As single-cell analysis has provided a better resolution to cell populations associated with poor clinical outcomes, the inventors analysed cell subpopulations for the enrichment of metastasis and treatment response associated with distinct gene expression signatures. For this, the inventors used 49 metastasis signature genes identified in patient-derived murine xenograft models of TNBC and which stratify high versus low metastatic burden tumors ³⁴. For therapy resistance signature genes, the inventors used 143 genes (from Cluster AHI) which were earlier found to be highly correlated with high Ki-67 expression and achieved the highest chemotherapy resistance scores in basal-like breast cancer subtypes ³⁵.

Functional analysis of identified subpopulation and reproducible signature gene identification

The set of genes that shows distinct expression in the identified cluster was compared between all three scRNA-seq datasets to identify reproducible signature genes across the same cell types in different studies. These reproducible genes were future analyses for biological functional analysis using the ShinyGO gene ontology enrichment analysis tool. The expression profiling of signature genes was investigated across TCGA breast cancer subtypes to ensure these signature genes are TNBC specific. Additionally, these genes were also screened in large breast cancer cohorts in EMT-High and EMT-Low groups. For this, the inventors first retrieved the TCGA mRNA expression z-scores of breast cancer tumors from cBioPortal ³⁶ using the ‘cgdsr’ R package (version 1.3.0) and extracted TNBC tumors based on the ER, PR and Her2 status. Next, the EMT scores were calculated for each sample by subtracting the average expression z-scores of the 5 ‘epithelial’ markers (ESRP1 , ESRP2, OVOL1 , OVOL2, and CLDN3) from the average expression z-scores of the 7 mesenchymal’ markers (ZEB1 , ZEB2, SNAI2, TWIST 1 , TWIST2, VIM and FN1). The tumor samples were then classified as EMT-High (defined by EMT scores > highest 1/3) and EMT-Low (defined by EMT scores ≤ lowest 1/3) based on the calculated EMT scores ³⁷.

The expression levels of these markers were utilized and spearman correlation between seven mesenchymal and five epithelial marker genes was calculated using the ‘corrplot’ R package (version 0.84) with a significant coefficient (95% confidence level; P-value <0.05), to select markers which separated patients with greater distinction. The ‘Pheatmap’ R package (version 1.0.12) was used to construct the heatmap of expression levels of the twelve markers in the EMT-High and EMT-Low groups defined above. Finally, the average expression profile of identified TFs in these EMT group's breast cancer cohorts was plotted using the ggplot2 package.

Development of RNA classifiers for predicting pathological complete response (pCR) or residual disease (RD)

For developing predictive models for stratifying TNBC tumours into pathological complete response (pCR) or residual disease (RD) groups, the inventors used four publicly available microarray-based gene expression datasets GSE25055 ³⁸, GSE25065 ³⁸, GSE20194 ³⁹, and GSE20271 ⁴⁰ which includes TNBC tumours. The inventors extracted normalised expression levels of 101 EMT signature genes from two studies of 180 patients to train the model. For testing the performance of the model the inventors used 130 TNBC patients from the remaining two datasets. The treatment details of all the patients were available and hence used as a class for binary classification. The inventors used Lasso and Elastic-Net Regularized Generalized Linear Models in glmnet for best fitting the model with 10-fold cross- validation to remove bias. In this, the inventors first ranked the features to select the minimum set of features with maximum predictive power for pCR and RD. The inventors next tested the model performance on independent validation datasets of 130 TNBC patients.

RESULTS

Metastasis and therapy resistance features associate with specific cell subpopulations within TNBC tumors

Here the inventors began by analyzing single-cell expression profiles of more than 10000 cells from 33 tumors of four independent TNBC studies (Fig. 1A). Further analysis of these datasets showed heterogeneous populations of immune, luminal, progenitor, and basal epithelial as highly enriched cell types which are defined by consistent expression of cell-type markers (Fig. 1 B-C; Fig. 6A). The inventors next investigated whether any of the identified subpopulations have enrichment of signature genes associated with aggressive clinical behaviour ³⁴’³⁵. An analysis of 49 signature genes shown to be associated with the metastatic burden (Patient-derived xenografts (PDX) of TNBC) ³⁴ and 143 genes highly activated in residual tumors of NAC-treated TNBC patients ³⁵ in the inventors’ data identified basal epithelial cells that exhibit high levels of both of these signature gene sets (Fig.lD, 6C-D). The inventors next extracted genes (n=101) that constitute these subpopulations reproducibly in at least two TNBC scRNA-seq datasets and consider as “signature genes” for further downstream analysis (Fig 1 E). The Kaplan-Meier survival analysis of these signature genes showed a significant decrease in 5-year relapse-free survival in more than 400 TNBC patients (Fig. 1F). These findings suggest that the aggressive features of the TNBC tumours, notably metastasis and therapy resistance, potentially emanate from small cell subpopulations within these heterogeneous tumors and contribute to a poor prognosis.

Pre-existence of specific active pathways associates with chemoresistance in TNBC To further confirm the role of identified signature genes in TNBC, the inventors analysed an existing scRNA-seq dataset of matched pre- and post- treatment TNBC dataset of four patients who responded (clonal extinction, chemosensitive) to neoadjuvant chemotherapy (NAC), docetaxel, and epirubicin versus four patients that persisted with a residual disease (clones persisted, chemoresistance) ¹⁵. Clustering of cells from both group patients revealed 13 clusters in each patient (Fig. 6D). Notably, patients from clonal extinct and persistent groups exhibited completely distinct clusters of pre and post treatment (Fig. 2A). Interestingly, profiling of signature genes in these patients groups revealed a significantly higher expression (p<2.22e-16) in clusters from chemoresistant patients (Fig. 2B). Given the higher expression of signature genes in chemoresistance patients, the inventors further identified genes that showed transcriptional reprogramming upon chemotherapy treatment. This results in a total of 24 genes which were showing upregulation upon chemotherapy treatment in chemoresistance patients and have no expression levels in chemosensitive patients (Fig. 2C).

The inventors were next interested in uncovering the transcription factor (TF) network potentially controlling the expression of signature genes underlying aggressive cell subpopulations. Towards this, the iRegulon tool was used to examine the enrichment of TF binding motifs at genomic loci encoding of the inventors’ signature genes. The analysis identified a strong enrichment of 26 transcription factors (Fig 2D) with SRF, MYLK and EP300 being the top enriched TFs that potentially regulate 64, 59 and 51 targets respectively among the signature genes (Fig 6D). Notably, the top 5 TFs (SRF, MYLK, EP300, ELF1 and NFIC) had highly overlapping sets of target genes suggesting their co-regulation of the gene expression program defining aggressive cell populations in TNBC (Fig 6D).

To further delineate functional pathways operating in these subpopulations associated with aggressive clinical behaviours, the inventors investigated transcriptional signatures of these cells using previously published hallmark signatures ⁴¹. This revealed enrichment of distinct categories of pathways in chemoresistance versus chemosensitive subpopulations (Fig. 2E). Interestingly, Epithelial to Mesenchymal Transition was among the pathways activated in chemoresistance subpopulations (highlighted with square box in Fig. 2E). EMT has been known to confer chemoresistance and linked aggressive behaviour in many cancers including TNBC ^6-11. In addition, the inventors also noticed AKT1 signalling in this category which has been associated with worse survival in TNBC ¹⁵ and implicated in paclitaxel resistance by inhibiting apoptotic pathways ¹⁵.

Given the limitations of the scRNA-seq technique in detecting lowly expression genes and to assess the robustness of these expression patterns in pooled RNA-seq data, the inventors further performed this analysis in a recently published bulk RNA- seq of 75 TNBC patients before (pre), during (mid) and after treatment with NAC (AC Adriamycin (Doxorubicin) + Cyclophosphamide, T Taxol (Docetaxel), H Herceptin (Trastuzumab) with known treatment outcome, namely pCR and RD status⁴² . In support of the inventors’ previous observations, the inventors’ signature genes showed significantly higher expression levels in patients with residual disease (RD) (Fig. 2F). These observations suggest that the induction of the inventors’ signature genes in response to NAC treatment may play a role in acquiring chemo-resistance in TNBC.

The cell subpopulations associated with TNBC aggressiveness are enriched with EMT features

Prompted by the inventors’ earlier observations, the inventors next attempted to further functionally characterize signature genes that define the identified aggressive subpopulations. A gene ontology (GO) analysis shows enrichment for EMT-like processes such as wound healing, extracellular matrix organization and cell migration (Fig. 3A). This is in line with previous observations where EMT was proposed to underlie metastasis and therapy resistance in TNBC. Furthermore, these genes were significantly highly expressed only in EMT-like subpopulations in all primary TNBC scRNA-seq datasets (Fig. 6A-B). Further analysis of a large TCGA breast cancer cohort showed that these genes are expressed at significantly higher levels in TNBC tumours compared to Luminal and Her2 breast cancers (Fig. 3B). The inventors next classified TCGA TNBC cancer samples into EMT-low and EMT- High groups based on the expression of 12 established markers (Fig. 3C) and then investigated the expression of the inventors’ signature genes. Interestingly, the inventors’ signature genes were observed to be significantly much higher expressed in patients of EMT-High groups compared to EMT-Low groups (Fig. 3D).

To better characterize these genes for EMT, the inventors induced EMT in immortalized human mammary epithelial cells (HMLEs) using TGFβ and performed transcriptome sequencing at different EMT time points representing early, mid and late EMT. Interestingly, expression analysis showed transcriptional induction of almost all signature genes (96 out of 101) during EMT, confirming that these genes are truly associated with mammary EMT (Fig. 3E). Interestingly however, distinct subsets of these genes showed induction at different time points during EMT, suggesting a potential division of labour during the cascade of EMT Progression.

Development of best in class prognostic R A-based classifiers for pCR and RD in TNBC

The pathological complete response (pCR) has been reported to be the key surrogate marker for long-term prognoses such as disease-free survival and overall survival in patients with triple-negative breast cancer ⁵’⁴³’⁴⁴. Previous research attempting to reveal predictors of pCR lack in achieving necessary predictive values for clinical utility for TNBC due to multiple factors such as small sample sizes, insufficient validation data or inapplicability to TNBC ³⁸'⁴⁰'⁴⁵-⁵². Importantly, while there are existing molecular tests to guide treatment for estrogen receptor (ER) or HER2+ tumors, no such tests exist in clinic to stratify pCR and residual disease (RD) in response to chemotherapy in TNBC. To address this unmet need, the inventors attempted to develop a predictive model which can predict pCR or RD to standard Neoadjuvant chemotherapy (NAC) in TNBC using expression of the inventors’ 101 signature genes in 307 TNBC patients from four breast cancer datasets (GSE25055, GSE25065, GSE20271 , GSE20194). These patients were treated with standard NAC incorporating taxane, anthracycline, and cyclophosphamide (AC-T), or additionally 5-fluorouracil (T-FAC), and their treatment outcome details (pCR and RD) were known.

For building a classifier of pCR and RD groups, the inventors first used Lasso and Elastic-Net Regularized Generalized Linear Models (glmnet) on 101 signature genes to reduce to a smaller set of genes with the highest predictive power. The inventors used the 10-fold cross-validation method to evaluate the discrimination ability to obtain a relatively unbiased estimate. Subsequently, a predictive model based on 20 genes was used to fit a generalized linear model (Fig. 4A, B, C). The 20 genes are as listed in Figure 4C. Next, the degree of discrimination measured using the receiver operating characteristic curve (ROC curve) showed that the predictive power of the inventors’ model based on 20 genes was highly accurate that achieved the area under the curve (AUC) of 0.903 (90.3%) (Fig. 4D). These 20 genes had a strong combinatorial accuracy as the inventors observed a decrease in the discriminative power if single genes or a set of genes were removed from the model (Fig. 4E, I). The inventors further compared the predictive efficiency of the inventors’ model with five published signature genes for predicting chemotherapy response^26-30. The inventors’ 20 gene gene panel (QUB) performed best compared to all existing panels in classifying pCR and RD in TNBC patients while requiring only 20 genes (Fig. 4F). Furthermore, this high performance was also valid for lymph node-positive TNBC samples (Fig. 4G). The Kaplan-Meier survival analysis showed that higher expression of these genes significantly reduces 5-year survival in TNBC patients (Fig. 4H top panel), including patients with advanced disease phenotype (lymph node-positive) (Fig. 4H bottom panel). The inventors next analysed the effect of removal of between 1 to 5 genes of the 20 genes on the performance of the model. The results are shown in Fig. 4I. ROC curves were generated for gene panels after removal of one to five of the top ranked genes (top panel); and after removal of one to five of the lower ranked genes (bottom panel). The gene(s) recited above each ROC curve is/are the gene(s) removed from the gene panel for the assessment; e.g, in the top panel, the first ROC curve is for a 19 gene gene panel consisting of the genes shown in the list on the left excluding ITGB1 , the second ROC curve is for an 18 gene gene panel consisting of the genes shown in the list on the left excluding ITGB1 and RBFOX2, etc. The results show that gene panels in which one to five of the of the five genes having greatest effect on the performance of the gene panel (i.e. one to five of ITGB1 , RBFOX2, DST, RCAN1 , c9orf3) are not included, the overall performance of the gene signature is still significantly superior to that of known gene signatures. The lower panel demonstrates that removal of the five genes having the least effect on the performance of the gene panel does not significantly reduce the performance of the gene panel, These findings suggest that the inventors’ gene panels outperform all current signatures in classifying pCR and RD in TNBC and provides the first potential test to facilitate upfront stratification of TNCB patients.

DISCUSSION

The aggressiveness in Triple-negative breast cancer arises due to the extensive intra-tumoral heterogeneity and presence of small subpopulations associated with chemoresistance and metastasis features. Here the inventors provided new insights into the contribution of cellular diversity in disease aggressiveness using integrated single-cell transcriptome analysis of more than 10000 cells from 33 TNBC patients. The inventors’ analysis of treatment-naive primary TNBC tumours¹⁴’^{31 32} revealed the existence of subpopulations associated with chemoresistance and metastasis. The inventors further showed that these subpopulations exhibit shared transcriptional profiles across all primary tumors of pre-treated TNBC patients (named as signature genes). Notably, the inventors discovered that differential levels of the same signature genes mark chemotherapy resistance TNBC tumors¹⁵. Furthermore, a differential expression of these signature genes was strongly linked with the relapse of the disease in TNBC patients within 5 years of the diagnosis. Concordant with recent studies, these findings strongly support that therapy-resistance cells are preexisted in subclones of treatment-naive patients, and further evolve to confer therapy resistance after chemotherapy treatment ¹⁵⁵³. Interestingly, the gene signatures defining these chemoresistance cell populations are involved in EMT-related processes. The role of EMT in disease aggressiveness such as chemoresistance and metastasis has been demonstrated in breast cancers including in TNBC ¹⁵’³⁴’^{54- 61}. For example, Kim et al showed that gene signatures related to EMT are upregulated in the chemoresistant tumor cells after chemotherapy treatment ¹⁵. The study by Zheng et al has demonstrated that mesenchymal cells in post-treatment tumours desensitize tumours to cytotoxic agents⁶². Similarly, a recent study using scRNA-seq profiling of migrating breast cancer cells found consistent activation of EMT and cancer stem cells (CSCs) related genes ⁶³. Furthermore, the inventors show that chemoresistant defining genes are co-regulated by known EMT- associated transcription factors such as SRF ⁶⁴⁶⁵, MYLK ⁶⁶, and ELF1 ⁶⁷⁶⁸ that play a key role in carcinogenesis, tumor progression, and drug resistance in several cancer types. Collectively, the inventors’ findings imply that the pre-existing chemoresistance phenotype is defined by a set of EMT-associated genes, which undergo transcriptional reprogramming in response to chemotherapy treatment and confer resistance in TNBC. This raises the possibility that targeting EMT signaling may re-sensitize the tumour cells to chemotherapy and will open new avenues in designing therapeutic strategies to overcome chemoresistance in TNBC ⁶⁹.

The discovered signature genes defining chemoresistance phenotype are highly robust and detected in bulk RNA-seq data of patients with residual disease and show greater differences in levels in TNBC tumors with high EMT status. In addition, with very few exceptions, almost all of these genes are upregulated at different time points during TGFβ-induced EMT in human mammary epithelial cells, suggesting a division of labour and the critical role of these genes in this process. These results are consistent with previous findings where expression of mesenchymal markers was highly evident in TNBC⁷⁰'⁷⁷.

Chemoresistance is a major challenge in TNBC management and no clinical signatures are available for this in clinical settings ²⁴'²⁶'⁷⁸-⁸⁰. Hence, the inventors’ 20 gene signature, that outperformed published signatures for chemotherapy prediction in TNBC²⁴’²⁶’⁷⁸’⁸⁰'⁸², holds strong clinical potential, notably, as they are derived from chemoresistant cell populations, highly reproducible across various single-cell and bulk transcriptomes, validated in more than 300 TNBC patients and strongly associated with poor survival. Predicting which patients will have pCR or RD using the inventors’ gene panel will provide an immense opportunity for clinicians to improve or consider alternative treatment plans and prevent loss of time due to unnecessary treatments and toxicity in TNBC.

In conclusion, the inventors’ study demonstrated the power of the single-cell analysis methods in uncovering mechanisms underlying aggressive clinical behaviour in TNBC particularly chemoresistance. The inventors’ study shows that chemoresistance cells pre-exit in the treatment of naive TNBC tumors and confer resistance phenotype via activated EMT programs when exposed to chemotherapy. Furthermore, the inventors’ highly accurate 20 gene signature model will impact the early evaluation of the effectiveness of systemic therapy and ensure long-term benefits in TNBC including later-stage tumors. In addition, a set of these signature genes show better sensitivity against a list of anticancer drugs, opening possibilities of alternate therapies. Overall, the inventors’ integrated analysis has discovered cell populations, underlying genes signatures, and mechanisms involved in TNBC chemoresistance and will not only facilitate therapy management by predicting therapy response but also create new therapeutic avenues for better clinical management of TNBC, thereby an improved survival and quality of life of these patients.

Although the invention has been particularly shown and described with reference to particular examples, it will be understood by those skilled in the art that various changes in the form and details may be made therein without departing from the scope of the present invention.

REFERENCES

1 Bianchini, G., Balko, J. M., Mayer, I. A., Sanders, M. E. & Gianni, L. Triplenegative breast cancer: challenges and opportunities of a heterogeneous disease. Nat Rev Clin Oncol 13, 674-690, doi:10.1038/nrclinonc.2016.66 (2016).

2 Malorni, L. et al. Clinical and biologic features of triple-negative breast cancers in a large cohort of patients with long-term follow-up. Breast Cancer Res Treat 136, 795-804, doi:10.1007/s10549-012-2315-y (2012).

3 Charpentier, M. & Martin, S. Interplay of Stem Cell Characteristics, EMT, and Microtentacles in Circulating Breast Tumor Cells. Cancers (Basel) 5, 1545- 1565, doi:10.3390/cancers5041545 (2013).

4 Foulkes, W. D., Smith, I. E. & Reis-Filho, J. S. Triple-negative breast cancer. N Engl J Med 363, 1938-1948, doi:10.1056/NEJMra1001389 (2010).

5 Liedtke, C. et al. Response to neoadjuvant therapy and long-term survival in patients with triple-negative breast cancer. J Clin Oncol 26, 1275-1281, doi:10.1200/JCG.2007.14.4147 (2008).

6 Jang, M. H., Kim, H. J., Kim, E. J., Chung, Y. R. & Park, S. Y. Expression of epithelial-mesenchymal transition-related markers in triple-negative breast cancer: ZEB1 as a potential biomarker for poor clinical outcome. Hum Pathol 46, 1267-1274, doi:10.1016/j.humpath.2015.05.010 (2015).

7 Bhola, N. E. et al. TGF-beta inhibition enhances chemotherapy action against triple-negative breast cancer. J Clin Invest 123, 1348-1358, doi:10.1172/JCI65416 (2013).

8 Xu, X. et al. TGF-beta plays a vital role in triple-negative breast cancer

(TNBC) drug-resistance through regulating sternness, EMT and apoptosis. Biochem Biophys Res Commun 502, 160-165, doi:10.1016/j.bbrc.2018.05.139 (2018). 9 Creighton, C. J. et al. Residual breast cancers after conventional therapy display mesenchymal as well as tumor-initiating features. Proc Natl Acad Sci U S A 106, 13820-13825, doi:10.1073/pnas.0905718106 (2009).

10 Luo, M., Brooks, M. & Wicha, M. S. Epithelial-mesenchymal plasticity of breast cancer stem cells: implications for metastasis and therapeutic resistance. Curr Pharm Des 21 , 1301-1310, doi:10.2174/1381612821666141211120604 (2015).

11 Hong, D. et al. Epithelial-to-mesenchymal transition and cancer stem cells contribute to breast cancer heterogeneity. J Cell Physiol 233, 9136-9144, doi:10.1002/jcp.26847 (2018).

12 Felipe Lima, J., Nofech-Mozes, S., Bayani, J. & Bartlett, J. M. EMT in Breast Carcinoma-A Review. J Clin Med 5, doi:10.3390/jcm5070065 (2016).

13 Koren, S. & Bentires-Alj, M. Breast Tumor Heterogeneity: Source of Fitness, Hurdle for Therapy. Mol Cell 60, 537-546, doi:10.1016/j.molcel.2015.10.031 (2015).

14 Karaayvaz, M. et al. Unravelling subclonal heterogeneity and aggressive disease states in TNBC through single-cell RNA-seq. Nat Commun 9, 3588, doi : 10.1038/S41467-018-06052-0 (2018).

15 Kim, C. et al. Chemoresistance Evolution in Triple-Negative Breast Cancer Delineated by Single-Cell Sequencing. Cell 173, 879-893 e813, doi: 10.1016Zj.cell.2018.03.041 (2018).

16 Shah, S. P. et al. The clonal and mutational evolution spectrum of primary triple-negative breast cancers. Nature 486, 395-399, doi: 10.1038/naturel 0933 (2012).

17 Turner, N. C. & Reis-Filho, J. S. Genetic heterogeneity and cancer drug resistance. Lancet Oncol 13, e178-185, doi:10.1016/S1470-2045(11)70335-7 (2012).

18 Houssami, N., Macaskill, P., Balleine, R. L., Bilous, M. & Pegram, M. D. HER2 discordance between primary breast cancer and its paired metastasis: tumor biology or test artefact? Insights through meta-analysis. Breast Cancer Res Treat V2S, 659-674, doi: 10.1007/s10549-011-1632-x (2011).

19 Bear, H. D. et al. Using the 21-gene assay from core needle biopsies to choose neoadjuvant therapy for breast cancer: A multicenter trial. J Surg Onco/ 115, 917-923, doi:10.1002/jso.24610 (2017).

20 Gianni, L. et al. Gene expression profiles in paraffin-embedded core biopsy tissue predict response to chemotherapy in women with locally advanced breast cancer. J Clin Oncol 23, 7265-7277, doi:10.1200/JC0.2005.02.0818 (2005).

21 Yardley, D. A. et al. A phase II trial of ixabepilone and cyclophosphamide as neoadjuvant therapy for patients with HER2-negative breast cancer: correlation of pathologic complete response with the 21-gene recurrence score. Breast Cancer Res Treat 154, 299-308, doi: 10.1007/s10549-015- 3613-y (2015).

22 Bertucci, F., Finetti, P., Viens, P. & Birnbaum, D. EndoPredict predicts for the response to neoadjuvant chemotherapy in ER-positive, HER2-negative breast cancer. Cancer Lett 355, 70-75, doi:10.1016/j.canlet.2014.09.014 (2014).

23 Prat, A. et al. Prediction of Response to Neoadjuvant Chemotherapy Using Core Needle Biopsy Samples with the Prosigna Assay. Clin Cancer Res 22, 560-566, doi: 10.1158/1078-0432.CCR-15-0630 (2016). 24 Zhao, Y., Schaafsma, E. & Cheng, C. Gene signature-based prediction of triple-negative breast cancer patient response to Neoadjuvant chemotherapy. Cancer Med 9, 6281-6295, doi:10.1002/cam4.3284 (2020).

25 Farmer, P. et al. A stroma-related gene signature predicts resistance to neoadjuvant chemotherapy in breast cancer. Nat Med 15, 68-74, doi: 10.1038/nm.1908 (2009)

26 Witkiewicz, A. K., Balaji, II. & Knudsen, E. S. Systematically defining singlegene determinants of response to neoadjuvant chemotherapy reveals specific biomarkers. Clin Cancer Res 20, 4837-4848, doi: 10.1158/1078-0432. CCR- 14-0885 (2014).

27 Juul, N. et al. Assessment of an RNA interference screen-derived mitotic and ceramide pathway metagene as a predictor of response to neoadjuvant paclitaxel for primary triple-negative breast cancer: a retrospective analysis of five clinical trials. Lancet Onco/ 11, 358-365, doi:10.1016/S1470-2045(10)70018-8 (2010).

28 Stover, D. G. et al. The Role of Proliferation in Determining Response to Neoadjuvant Chemotherapy in Breast Cancer: A Gene Expression-Based Meta-Analysis. Clin Cancer Res 22, 6039-6050, doi: 10.1158/1078- 0432.CCR-16-0471 (2016).

29 Fournier, M. V. et al. A Predictor of Pathological Complete Response to Neoadjuvant Chemotherapy Stratifies Triple Negative Breast Cancer Patients with High Risk of Recurrence. Sci Rep 9, 14863, doi: 10.1038/s41598-019- 51335-1 (2019).

30 Lim, G. B. et al. Prediction of prognostic signatures in triple-negative breast cancer based on the differential expression analysis via NanoString nCounter immune panel. BMC Cancer 20, 1052, doi:10.1186/s12885-020-07399-8 (2020).

31 Chung, W. et al. Single-cell RNA-seq enables comprehensive tumour and immune cell profiling in primary breast cancer. Nat Commun 8, 15081, doi: 10.1038/ncommsl 5081 (2017).

32 Gulati, G. S. et al. Single-cell transcriptional diversity is a hallmark of developmental potential. Science 367, 405-411, doi : 10.1126/science. aax0249 (2020) .

33 Zhang, X. et al. Cel I Marker: a manually curated resource of cell markers in human and mouse. Nucleic Acids Res 47, D721-D728, doi:10.1093/nar/gky900 (2019).

34 Lawson, D. A. et al. Single-cell analysis reveals a stem-cell program in human metastatic breast cancer cells. Nature 526, 131-135, doi: 10.1038/naturel 5260 (2015).

35 Balko, J. M. et al. Profiling of residual breast cancers after neoadjuvant chemotherapy identifies DLISP4 deficiency as a mechanism of drug resistance. Nat Med 18, 1052-1059, doi:10.1038/nm.2795 (2012).

36 Gao, J. et al. Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci Signal 6, pl1 , doi:10.1126/scisignal.2004088 (2013).

37 Lou, Y. et al. Epithelial-Mesenchymal Transition Is Associated with a Distinct Tumor Microenvironment Including Elevation of Inflammatory Signals and Multiple Immune Checkpoints in Lung Adenocarcinoma. Clin Cancer Res 22, 3630-3642, doi:10.1158/1078-0432.CCR-15-1434 (2016).

38 Hatzis, C. et al. A genomic predictor of response and survival following taxane-anthracycline chemotherapy for invasive breast cancer. JAMA 305, 1873-1881 , doi:10.1001/jama.2011.593 (2011). 39 Shi, L. et al. The MicroArray Quality Control (MAQC)-II study of common practices for the development and validation of microarray-based predictive models. Nat Biotechnol 28, 827-838, doi:10.1038/nbt.1665 (2010).

40 Tabchy, A. et al. Evaluation of a 30-gene paclitaxel, fluorouracil, doxorubicin, and cyclophosphamide chemotherapy response predictor in a multicenter randomized trial in breast cancer. Clin Cancer Res 16, 5351-5361, doi: 10.1158/1078-0432.CCR-10-1265 (2010).

41 Liberzon, A. et al. The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst 1, 417-425, doi: 10.1016/j. cels.2015.12.004 (2015).

42 Park, Y. H. et al. Chemotherapy induces dynamic immune responses in breast cancers that impact treatment outcome. Nat Commun 11, 6175, doi : 10.1038/S41467-020- 19933-0 (2020) .

43 von Minckwitz, G. et al. Definition and impact of pathologic complete response on prognosis after neoadjuvant chemotherapy in various intrinsic breast cancer subtypes. J Clin Oncol 30, 1796-1804, doi:10.1200/JC0.2011.38.8595 (2012).

44 Hahnen, E. et al. Germline Mutation Status, Pathological Complete Response, and Disease-Free Survival in Triple-Negative Breast Cancer: Secondary Analysis of the GeparSixto Randomized Clinical Trial. JAMA Oncol 3, 1378-1385, doi:10.1001/jamaoncol.2017.1007 (2017).

45 Ayers, M. et al. Gene expression profiles predict complete pathologic response to neoadjuvant paclitaxel and fluorouracil, doxorubicin, and cyclophosphamide chemotherapy in breast cancer. J Clin Oncol 22, 2284- 2293, doi:10.1200/JCQ.2004.05.166 (2004).

46 Lehmann, B. D. et al. Refinement of Triple-Negative Breast Cancer Molecular Subtypes: Implications for Neoadjuvant Chemotherapy Selection. PLoS One 11 , e0157368, doi:10.1371/journal.pone.0157368 (2016).

47 Louie, M. C. & Sevigny, M. B. Steroid hormone receptors as prognostic markers in breast cancer. Am J Cancer Res 7, 1617-1636 (2017).

48 Mark, K. M. K., Varn, F. S., Ung, M. H., Qian, F. & Cheng, C. The E2F4 prognostic signature predicts pathological response to neoadjuvant chemotherapy in breast cancer patients. BMC Cancer 17, 306, doi: 10.1186/s 12885-017-3297-2 (2017).

49 Masuda, H. et al. Differential response to neoadjuvant chemotherapy among 7 triple-negative breast cancer molecular subtypes. Clin Cancer Res 19, 5533-5540, doi: 10.1158/1078-0432.CCR-13-0799 (2013).

50 Nakashoji, A. et al. Clinical predictors of pathological complete response to neoadjuvant chemotherapy in triple-negative breast cancer. Oncol Lett 14, 4135-4141, doi:10.3892/ol.2017.6692 (2017).

51 Nwaogu, I. Y., Fayanju, O. M., Jeffe, D. B. & Margenthaler, J. A. Predictors of pathological complete response to neoadjuvant chemotherapy in stage II and III breast cancer: The impact of chemotherapeutic regimen. Mol Clin Oncol 3, 1117-1122, doi:10.3892/mco.2015.579 (2015).

52 Santuario-Facio, S. K. et al. A New Gene Expression Signature for Triple Negative Breast Cancer Using Frozen Fresh Tissue before Neoadjuvant Chemotherapy. Mol Med 23, 101-111, doi:10.2119/molmed.2016.00257 (2017).

53 Brady, S. W. et al. Combating subclonal evolution of resistant cancer phenotypes. Nat Commun 8, 1231, doi: 10.1038/s41467-017-01174-3 (2017). 54 Liao, T. T. & Yang, M. H. Revisiting epithelial-mesenchymal transition in cancer metastasis: the connection between epithelial plasticity and sternness. Mol Onco/ 11, 792-804, doi:10.1002/1878-0261.12096 (2017).

55 Mani, S. A. et al. The epithelial-mesenchymal transition generates cells with properties of stem cells. Cell 133, 704-715, doi: 10.1016/j. cell.2008.03.027 (2008).

56 Blanco, M. J. et al. Correlation of Snail expression with histological grade and lymph node status in breast carcinomas. Oncogene 21 , 3241-3246, doi:10.1038/sj.onc.1205416 (2002).

57 Li, X. et al. Intrinsic resistance of tumorigenic breast cancer cells to chemotherapy. J Natl Cancer Inst 100, 672-679, doi: 10.1093/jnci/djn123 (2008).

58 Taube, J. H. et al. Core epithelial-to-mesenchymal transition interactome gene-expression signature is associated with claudin-low and metaplastic breast cancer subtypes. Proc Natl Acad Sci U S A 107, 15449-15454, doi : 10.1073/pnas.1004900107 (2010) .

59 Tian, M. & Schiemann, W. P. TGF-beta Stimulation of EMT Programs Elicits Non-genomic ER-alpha Activity and Anti-estrogen Resistance in Breast Cancer Cells. J Cancer Metastasis Treat 3, 150-160, doi: 10.20517/2394- 4722.2017.38 (2017).

60 Oliveras-Ferraros, C. et al. Epithelial-to-mesenchymal transition (EMT) confers primary resistance to trastuzumab (Herceptin). Cell Cycle 11, 4020- 4032, doi:10.4161/cc.22225 (2012).

61 Li, X., Strietz, J., Bleilevens, A., Stickeler, E. & Maurer, J. Chemotherapeutic Stress Influences Epithelial-Mesenchymal Transition and Sternness in Cancer Stem Cells of Triple-Negative Breast Cancer. Int J Mol Sci 21, doi:10.3390/ijms21020404 (2020).

62 Zheng, X. et al. Epithelial-to-mesenchymal transition is dispensable for metastasis but induces chemoresistance in pancreatic cancer. Nature 527, 525-530, doi: 10.1038/naturel 6064 (2015).

63 Chen, Y. C. et al. Single-cell RNA-sequencing of migratory breast cancer cells: discovering genes associated with cancer metastasis. Analyst 144, 7296-7309, doi:10.1039/c9an01358j (2019).

64 Zhao, X. et al. SRF expedites metastasis and modulates the epithelial to mesenchymal transition by regulating miR-199a-5p expression in human gastric cancer. Cell Death Differ 21 , 1900-1913, doi:10.1038/cdd.2014.109 (2014).

65 Bae, J. S. et al. Serum response factor induces epithelial to mesenchymal transition with resistance to sorafenib in hepatocellular carcinoma. Int J Oncol 44, 129-136, doi: 10.3892/ijo.2013.2154 (2014).

66 Sundararajan, V. et al. The ZEB1/miR-200c feedback loop regulates invasion via actin interacting proteins MYLK and TKS5. Oncotarget 6, 27083-27096, doi:10.18632/oncotarget.4807 (2015).

67 Budka, J. A., Ferris, M. W., Capone, M. J. & Hollenhorst, P. C. Common ELF1 deletion in prostate cancer bolsters oncogenic ETS function, inhibits senescence and promotes docetaxel resistance. Genes Cancer 9, 198-214, doi: 10.18632/genesandcancer.182 (2018).

68 Wang, L. ELF1 -activated FOXD3-AS1 promotes the migration, invasion and EMT of osteosarcoma cells via sponging miR-296-5p to upregulate ZCCHC3. J Bone Oncol 26, 100335, doi: 10.1016/j.jbo.2020.100335 (2021). 69 Marcucci, F., Stassi, G. & De Maria, R. Epithelial-mesenchymal transition: a new target in anticancer drug discovery. Nat Rev Drug Discov 15, 311-325, doi:10.1038/nrd.2015.13 (2016).

70 Yang, J. H. et al. Snail augments fatty acid oxidation by suppression of mitochondrial ACC2 during cancer progression. Life Sci Alliance 3, doi:10.26508/lsa.202000683 (2020).

71 Vijay, G. V. et al. GSK3beta regulates epithelial-mesenchymal transition and cancer stem cell properties in triple-negative breast cancer. Breast Cancer Res 21 , 37, doi:10.1186/s13058-019-1125-0 (2019).

72 Prat, A. et al. Phenotypic and molecular characterization of the claudin-low intrinsic subtype of breast cancer. Breast Cancer Res 12, R68, doi:10.1186/bcr2635 (2010).

73 Pomp, V. et al. Differential expression of epithelial-mesenchymal transition and stem cell markers in intrinsic subtypes of breast cancer. Breast Cancer Res Treat 154, 45-55, doi:10.1007/s10549-015-3598-6 (2015).

74 Maturi, V., Moren, A., Enroth, S., Heldin, C. H. & Moustakas, A. Genomewide binding of transcription factor SnaiH in triple-negative breast cancer cells. Mol Oncol 12, 1153-1174, doi: 10.1002/1878-0261.12317 (2018).

75 Ito, K., Park, S. H., Nayak, A., Byerly, J. H. & Irie, H. Y. PTK6 Inhibition Suppresses Metastases of Triple-Negative Breast Cancer via SNAIL- Dependent E-Cadherin Regulation. Cancer Res 76, 4406-4417, doi: 10.1158/0008-5472.CAN-15-3445 (2016).

76 Sarrio, D. et al. Epithelial-mesenchymal transition in breast cancer relates to the basal-like phenotype. Cancer Res 68, 989-997, doi: 10.1158/0008- 5472.CAN-07-2017 (2008).

77 Trimboli, A. J. et al. Direct evidence for epithelial-mesenchymal transitions in breast cancer. Cancer Res 68, 937-945, doi: 10.1158/0008-5472. CAN-07- 2148 (2008).

78 Ignatiadis, M. et al. Gene modules and response to neoadjuvant chemotherapy in breast cancer subtypes: a pooled analysis. J Clin Oncol 30, 1996-2004, doi:10.1200/JCG.2011.39.5624 (2012).

79 Gingras, I., Desmedt, C., Ignatiadis, M. & Sotiriou, C. CCR 20th Anniversary Commentary: Gene-Expression Signature in Breast Cancer--Where Did It Start and Where Are We Now? Clin Cancer Res 21 , 4743-4746, doi : 10.1158/1078-0432. CC R- 14-3127 (2015) .

80 Pineda, B. et al. A two-gene epigenetic signature for the prediction of response to neoadjuvant chemotherapy in triple-negative breast cancer patients. Clin Epigenetics 11 , 33, doi: 10.1186/s13148-019-0626-0 (2019).

81 Denkert, C. et al. Tumor-infiltrating lymphocytes and response to neoadjuvant chemotherapy with or without carboplatin in human epidermal growth factor receptor 2-positive and triple-negative primary breast cancers. J Clin Oncol 33, 983-991, doi:10.1200/JC0.2014.58.1967 (2015).

82 Loi, S. et al. Tumor infiltrating lymphocytes are prognostic in triple negative breast cancer and predictive for trastuzumab benefit in early breast cancer: results from the FinHER trial. Ann Oncol 25, 1544-1550, doi:10.1093/annonc/mdu112 (2014).

Claims

Claims

1. A method of predicting resistance to chemotherapy in a subject having triple negative breast cancer (TNBC), wherein said method comprises: a) providing a biological sample from said subject; and b) determining the expression levels of each member of a gene panel, said gene panel comprising at least 10 genes selected from the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1, STOM, COL6A1, DSC3, and AMIGO2 in said biological sample; wherein the determined expression levels of said genes of said gene panel is used to determine the likelihood of resistance to said chemotherapy, wherein said gene panel comprises one, two, three, four, five or six of ITGB1 , RBFOX2, DST, RCAN1 , c9orf3 and ACTA2.

2. The method according to claim 1 , wherein said gene panel comprises at least 10 genes selected from the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, NFIB, and ACTN1.

3. The method according to claim 1 or claim 2, wherein the gene panel comprises at least 15 genes selected from the group consisting of ITGB1 , RBFOX2, DST, RCAN1 , c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1 , DSC3, and AMIGO2

4. The method according to claim 3, wherein said gene panel comprises at least the fifteen genes of the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, NFIB, and ACTN1.

5. The method according to claim 4, wherein said gene panel comprises (i) at least the fifteen genes of the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1 , PRNP, TIMP3, CD63, IFI16, NFIB, and ACTN1 ; and optionally (ii) one or more of the genes of the group consisting of SFRP1 , STOM, COL6A1, DSC3, and AMIGO2.

6. The method according to any one of the preceding claims wherein said gene panel comprises at least one, two or three genes selected from the group consisting of: c9orf3, CD63, and STOM.

7. The method according to claim 8 wherein said gene panel comprises c9orf3, CD63, and STOM

8. The method according to any one of the preceding claims wherein said gene panel comprises the twenty genes of the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1 , STOM, COL6A1 , DSC3, and AMIGO2.

9. The method according to any one of the preceding claims wherein the gene panel further comprises one or more of the genes selected from SLC11A2, MSH3, TSHZ2, CTDSPL, and NDUFA6.

10. The method according to any one of the preceding claims wherein the gene panel comprises no more than 25 genes.

11. The method according to claim 10 wherein the gene panel comprises no more than 20 genes.

12. The method according to claim 11 , wherein the gene panel comprises only genes selected from the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1, DSC3, AMIGO2, SLC11A2, MSH3, TSHZ2, CTDSPL, and NDUFA6.

13. The method according to claim 12, wherein the gene panel comprises only genes selected from the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1 , STOM, COL6A1 , DSC3, and AMIGO2. The method according to any one of claims 1 to 4, wherein the gene panel consists of:

(a) ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1, STOM, COL6A1, DSC3, and AMIGO2;

(b) RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1, STOM, COL6A1, DSC3, and AMIGO2

(c) DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1 , STOM, COL6A1 , DSC3, and AMIGO2;

(d) RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1 , STOM, COL6A1 , DSC3, and AMIGO2;

(e) c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1 , STOM, COL6A1, DSC3, and AMIGO2; or

(f) ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1, SFRP1 , STOM, COL6A1 , DSC3, and AMIGO2. A method of predicting resistance to chemotherapy in a subject having triple negative breast cancer (TNBC), wherein said method comprises: a) providing a biological sample from said subject; and b) determining the expression levels of each member of a gene panel, said gene panel comprising at least 10 genes selected from the group consisting of ITGB1, RBFOX2, DST, RCAN1, c9orf3, ACTA2, S100B, LY6E, CTNNAL1, PRNP, TIMP3, CD63, IFI16, NFIB, ACTN1 , SFRP1, STOM, COL6A1, DSC3, and AMIGO2 in said biological sample; wherein the determined expression levels of said genes of said gene panel is used to determine the likelihood of resistance to said chemotherapy. The method according to any one of the preceding claims, wherein said method comprises determining the expression levels of said genes of the gene panel relative to reference amounts of said genes. The method according to any one of the preceding claims wherein the likelihood of resistance to said chemotherapy is determined by applying the expression levels to a predictive model which relates expression levels of said genes of the gene panel with resistance to chemotherapy against triple negative breast cancer.

18. The method according to claim 17, wherein applying the expression levels to a predictive model comprises weighting the expression levels of said genes of the gene panel according to a predetermined ranking of said genes of the gene panel.

19. The method according to any one of the preceding claims wherein

(a) where the gene panel comprises one or more of the genes of the group consisting of RBFOX2, DST, RCAN1, c9orf3, ACTA2, LY6E, PRNP, TIMP3, CD63, ACTN1 , STOM, COL6A1, DSC3, and AMIGO2, expression of said one or more of said genes at a significantly higher level than in control cells is indicative of the sample being a sample of TNBC cells which are chemotherapy resistant; and/or

(b) where the gene panel comprises one or more of the genes of the group consisting of ITGB1, S100B, CTNNAL1, IFI16, NFIB, and SFRP1, expression of said one or more of said genes at a significantly lower level than in control cells is indicative of the sample being a sample of TNBC cells which are chemotherapy resistant.

20. The method according to any one of the preceding claims wherein the AUC associated with said gene panel is greater than 0.850.

21. The method according to any one of the preceding claims wherein the chemotherapy is combination chemotherapy comprising an anthracycline and an alkylating agent.

22. The method according to claim 21 , wherein the chemotherapy is combination chemotherapy which further comprises an antimetabolite.