US20100112603A1

US20100112603A1 - Method for predicting the response to a treatment

Info

Publication number: US20100112603A1
Application number: US12/624,443
Authority: US
Inventors: Joachim Moecks; Andreas Strauss; Gerhard Zugmaier
Original assignee: Individual
Current assignee: Individual
Priority date: 2005-08-12
Filing date: 2009-11-24
Publication date: 2010-05-06
Also published as: MA31549B1; HRP20110171T1; ES2356066T3; JP2009504142A; NO20080371L; CL2010000720A1; AU2011202339A1; KR20080034922A; NZ565270A; WO2007019899A2; AR057323A1; AU2006281746B2; CA2616913A1; PT1915460E; PL1915460T3; EP2196547B1; RU2010115252A; ECSP088171A; TW200741010A; ATE539171T1

Abstract

The invention is related to a method of predicting the response to a treatment with a HER inhibitor in a patient comprising the steps of assessing a biomarker or a combination of biomarkers selected from the group consisting of amphiregulin, an epidermal growth factor, a transforming growth factor alpha, and a HER2 biomarker in a biological sample from the patient and predicting the response to the treatment with the HER inhibitor in the patient by evaluating the results of the first step. Further uses and methods wherein these markers are used are disclosed.

Description

PRIORITY TO RELATED APPLICATIONS

This application is a continuation of U.S. application Ser. No. 11/438,033, filed May 19, 2006, now pending; which claims the benefit of European Application No. 05017663.5, filed Aug. 12, 2005. The entire contents of the above-identified applications are hereby incorporated by reference.

FIELD OF THE INVENTION

The invention is related to a method of predicting the response to a treatment with a HER inhibitor, preferably a HER dimerization inhibitor, in a patient comprising the steps of assessing a marker gene or a combination of marker genes selected from the group consisting of an epidermal growth factor, a transforming growth factor alpha and a HER2 marker gene or a combination of marker genes comprising an amphiregulin marker gene and a marker gene selected from an epidermal growth factor, a transforming growth factor alpha and a HER2 marker gene in a biological sample from the patient and predicting the response to the treatment with the HER inhibitor in the patient by evaluating the results of the first step. Further uses and methods wherein these markers are used are disclosed.

BACKGROUND OF THE INVENTION

The human epidermal growth factor receptor (ErbB or HER) family comprises four members (HER1-4) that, through the activation of a complex signal cascade, are important mediators of cell growth, survival and differentiation. At least 11 different gene products from the epidermal growth factor (EGF) superfamily bind to three of these receptors, EGFR (also called ErbB1 or HER1), HER3 (ErbB3) and HER4 (ErbB4). Although no ligand has been identified that binds and activates HER2 (ErbB2 or neu), the prevailing understanding is that HER2 is a co-receptor that acts in concert with other HER receptors to amplify and in some cases initiate receptor-ligand signaling. Dimerization with the same receptor type (homodimerization) or another member of the HER family (heterodimerization) is essential for their activity. HER2 is the preferred dimerization partner for other HER family members. The role of the HER family in many epithelial tumor types is well documented and has led to the rational development of novel cancer agents directed specifically to HER receptors. The recombinant humanized anti-HER2 monoclonal antibody (MAb) trastuzumab is a standard of care in patients with HER2-positive metastatic breast cancer (MBC). Overexpression/amplification of the HER2 protein/gene, which occurs in 20-30% of breast cancer cases, is a prerequisite for treatment with trastuzumab.
Pertuzumab (Omnitarg™; formerly 2C4) is the first of a new class of agents known as HER dimerization inhibitors (HDIs). Pertuzumab binds to HER2 at its dimerization domain, thereby inhibiting its ability to form active dimer receptor complexes and thus blocking the downstream signal cascade that ultimately results in cell growth and division. Pertuzumab is a fully humanized recombinant monoclonal antibody directed against the extracellular domain of HER2. Binding of Pertuzumab to the HER2 on human epithelial cells prevents HER2 from forming complexes with other members of the HER family (including EGFR, HER3, HER4) and probably also HER2 homodimerization. By blocking complex formation, Pertuzumab prevents the growth-stimulatory effects and cell survival signals activated by ligands of HER1, HER3 and HER4 (e.g. EGF, TGFα, amphiregulin, and the heregulins). Other names for Pertuzumab are 2C4 or Pertuzumab. Pertuzumab is a fully humanized recombinant monoclonal antibody based on the human IgG1(κ) framework sequences. The structure of Pertuzumab consists of two heavy chains (449 residues) and two light chains (214 residues). Compared to Trastuzumab (Herceptin®), Pertuzumab has 12 amino acid differences in the light chain and 29 amino acid differences in the IgG1 heavy chain. WO 2004/092353 and WO 2004/091384 present investigations that the formation of heterodimers of HER2 with other receptors should be linked to the effectiveness or suitability of Pertuzumab.
Zabrecky, J. R. et al., J. Biol. Chem. 266 (1991) 1716-1720 disclose that the release of the extracellular domain of HER2 may have implications in oncogenesis and its detection could be useful as a cancer diagnostic. Colomer, R. et al., Clin. Cancer Res. 6 (2000) 2356-2362 disclose circulating HER2 extracellular domain and resistance to chemotherapy in advanced breast cancer. The prognostic and predictive values of the extracellular domain of HER2 is reviewed by Hait, W. N., Clin. Cancer Res. 7 (2001) 2601-2604.

SUMMARY OF THE INVENTION

There is still a need to provide further methods for determining the progression of disease in a cancer patient treated with a HER dimerization inhibitor.
Therefore, in an embodiment of the invention, a method of predicting the response to a treatment with a HER inhibitor, preferably a HER dimerization inhibitor, in a patient is provided comprising the steps of:

- (a) determining the expression level or amount of one or more biomarker in a biological sample from a patient wherein the biomarker or biomarkers are selected from the group consisting of:
  - (1) transforming growth factor alpha;
  - (2) HER2;
  - (3) amphiregulin; and
  - (4) epidermal growth factor;
- (b) determining whether the expression level or amount assessed in step (a) is above or below a certain quantity that is associated with an increased or decreased clinical benefit to a patient; and
- (c) predicting the response to the treatment with the HER inhibitor in the patient by evaluating the results of step (b).

In another embodiment of the invention, a probe that hybridizes with the polynucleotides of the above biomarkers under stringent conditions or an antibody that binds to the proteins of the above biomarkers is used for predicting the response to treatment with a HER inhibitor in a patient or used for selecting a composition for inhibiting the progression of disease in a patient.
In still another embodiment of the invention, a kit is provided comprising a probe that anneals with a biomarker polynucleotide under stringent conditions or an antibody that binds to the biomarker protein.
In still another embodiment of the invention, a method of selecting a composition for inhibiting the progression of disease in a patient is provided, the method comprising:

- (a) separately exposing aliquots of a biological sample from a cancer patient in the presence of a plurality of test compositions;
- (b) comparing the level of expression of one or more biomarkers selected from the group consisting of amphiregulin, epidermal growth factor, transforming growth factor alpha and HER2 in the aliquots of the biological sample contacted with the test compositions and the level of expression of such biomarkers in an aliquot of the biological sample not contacted with the test compositions; and
- (c) selecting one of the test compositions which alters the level of expression of a particular biomarker or biomarkers in the aliquot of the biological sample contacted with the test composition and the level of expression of the corresponding biomarker or biomarkers in the aliquot of the biological sample not contacted with the test composition is an indication for the selection of the test composition.

In yet another embodiment of the invention, a method of identifying a candidate agent is provided said method comprising:

- (a) contacting an aliquot of a biological sample from a cancer patient with the candidate agent and determining the level of expression of one or more biomarkers selected from the group consisting of amphiregulin, epidermal growth factor, transforming growth factor alpha and HER2 in the aliquot;
- (b) determining the level of expression of a corresponding biomarker or of a corresponding combination of biomarkers in an aliquot of the biological sample not contacted with the candidate agent;
- (c) observing the effect of the candidate agent by comparing the level of expression of the biomarker or biomarkers in the aliquot of the biological sample contacted with the candidate agent and the level of expression of the corresponding biomarker or biomarkers in the aliquot of the biological sample not contacted with the candidate agent; and
- (d) identifying said agent from said observed effect, wherein an at least 10% difference between the level of expression of the biomarker or biomarkers in the aliquot of the biological sample contacted with the candidate agent and the level of expression of the corresponding biomarker or biomarkers in the aliquot of the biological sample not contacted with the candidate agent is an indication of an effect of the candidate agent.

In yet another embodiment, a candidate agent identified by the method according to the invention or a pharmaceutical preparation comprising an agent according to the invention is provided.
In yet another embodiment of the invention, an agent according to the invention is provided for the preparation of a composition for the treatment of cancer.
In still another embodiment of the invention, a method of producing a drug is provided comprising:

- (i) synthesizing the candidate agent identified as described above or an analog or derivative thereof in an amount sufficient to provide said drug in a therapeutically effective amount to a subject; and/or
- (ii) combining the drug candidate or the candidate agent identified as described above or an analog or derivative thereof with a pharmaceutically acceptable carrier.

In yet another embodiment of the invention, a biomarker protein or a biomarker polynucleotide selected from the group consisting of an amphiregulin biomarker, and epidermal growth factor biomarker, a transforming growth factor alpha biomarker and a HER2 biomarker protein or polynucleotide is used for deriving a candidate agent or for selecting a composition for inhibiting the progression of a disease in a patient.
In another embodiment of the invention, a HER inhibitor, preferably a HER dimerization inhibitor, is used for the manufacture of a medicament for treating a human cancer patient characterized in that said treating or treatment includes assessing in a biological sample from the patient: one or more biomarkers selected from the group consisting of amphiregulin biomarker, epidermal growth factor biomarker, transforming growth factor alpha biomarker, and HER2 biomarker. In a particular embodiment, one or more biomarkers are assessed wherein the biomarkers are selected from the group consisting of epidermal growth factor, transforming growth factor alpha, and HER2. In another particular embodiment, a transforming growth factor alpha biomarker is assessed in combination with one or more biomarkers selected from the group consisting of epidermal growth factor, amphiregulin, and HER2. In another particular embodiment, a HER2 biomarker is assessed in combination with one or more biomarkers selected from the group consisting of epidermal growth factor, transforming growth factor alpha, and amphiregulin.
In another particular embodiment, a epidermal growth factor biomarker is assessed in combination with one or more biomarkers selected from the group consisting of amphiregulin, transforming growth factor alpha, and HER2.
In another particular embodiment, an amphiregulin biomarker is assessed in combination with one or more biomarkers selected from the group consisting of epidermal growth factor, transforming growth factor alpha, and HER2.

BRIEF DESCRIPTION OF THE FIGURES

FIG. 1: Scatterplot TGF-alpha logarithmic transformation versus categorized clinical benefit

FIG. 2: Scatterplot Amphiregulin logarithmic transformation versus categorized clinical benefit

FIG. 3: Ordinal clinical benefit TGF-alpha

FIG. 4: Ordinal clinical benefit Amphiregulin

FIG. 5: Ordinal clinical benefit EGF

FIG. 6: Ordinal clinical benefit HER2-ECD

FIG. 7: Overview exploratory cut-points and log-rank p-values for TTP and TTD for Amphiregulin, EGF, TGF-alpha, HER2-ECD

FIG. 8: TGF-alpha Kaplan Meier plot for time to progression/death based on exploratory single marker cut-point

FIG. 9: TGF-alpha Kaplan Meier plot for time to death based on exploratory single marker cut-point

FIG. 10: Amphiregulin Kaplan Meier plot for time to progression/death based on exploratory single marker cut-point

FIG. 11: Amphiregulin Kaplan Meier plot for time to death based on exploratory single marker cut-point

FIG. 12: EGF Kaplan Meier plot for time to progression/death based on exploratory single marker cut-point

FIG. 13: EGF Kaplan Meier plot for time to death based on exploratory single marker cut-point

FIG. 14: HER2-ECD Kaplan Meier plot for time to progression/death based on exploratory single marker cut-point

FIG. 15: HER2-ECD Kaplan Meier plot for time to death based on exploratory single marker cut-point

FIG. 16: As example for a combination score, further improving the separation between the greater clinical benefit/lesser clinical benefit groups in TTP: Ordinal clinical benefit HER2-ECD TGF alpha combination

FIG. 17: Overview exploratory cut-points and log-rank p-values for TTP and TTD for a combination of TGF-alpha and HER2-ECD

FIG. 18: HER2-ECD/TGF-alpha Kaplan Meier plot for time to progression/death based on exploratory combination marker cut-point

FIG. 19: HER2-ECD/TGF-alpha Kaplan Meier plot for time to death based on exploratory combination marker cut-point

DETAILED DESCRIPTION OF THE INVENTION

The articles “a” and “an” are used herein to refer to one or to more than one (i.e. to at least one) of the grammatical object of the article. By way of example, “an element” means one element or more than one element.
The term “biological sample” shall generally mean any biological sample obtained from an individual, body fluid, cell line, tissue culture, or other source. Body fluids are e.g. lymph, sera, plasma, urine, semen, synovial fluid and spinal fluid. Methods for obtaining tissue biopsies and body fluids from mammals are well known in the art. If the term “sample” is used alone, it shall still mean that the “sample” is a “biological sample”, i.e. the terms are used interchangeably.
The term “response of a patient to treatment with a HER inhibitor” or “response of a patient to treatment with a HER dimerization inhibitor” refers to the clinical benefit imparted to a patient suffering from a disease or condition (such as cancer) from or as a result of the treatment with the HER inhibitor (e.g., a HER dimerization inhibitor). A clinical benefit includes a complete remission, a partial remission, a stable disease (without progression), progression-free survival, disease free survival, improvement in the time-to-progression (of the disease), improvement in the time-to-death, or improvement in the overall survival time of the patient from or as a result of the treatment with the HER dimerization inhibitor. There are criteria for determining a response to therapy and those criteria allow comparisons of the efficacy to alternative treatments (Slapak and Kufe, Principles of Cancer Therapy, in Harrisons's Principles of Internal Medicine, 13th edition, eds. Isselbacher et al., McGraw-Hill, Inc., 1994). For example, a complete response or complete remission of cancer is the disappearance of all detectable malignant disease. A partial response or partial remission of cancer may be, for example, an approximately 50 percent decrease in the product of the greatest perpendicular diameters of one or more lesions or where there is not an increase in the size of any lesion or the appearance of new lesions.
As used herein, the term “progression of cancer” includes and may refer to metastasis; a recurrence of cancer, or an at least approximately 25 percent increase in the product of the greatest perpendicular diameter of one lesion or the appearance of new lesions. The progression of cancer, preferably breast cancer, is “inhibited” if recurrence or metastasis of the cancer is reduced, slowed, delayed, or prevented.
As used herein, the term “Time To Progression/death” (also referred to as “TPP”) or Progression-Free Survival (also referred to as “PFS”) refers to a clinical endpoint frequently used in oncology trials (that includes but is not limited to clinical trials with reference to the present invention). The measurement for each patient equals the time elapsed from onset of the treatment of a patient in a trial (as defined in the protocol [i.e, see the examples infra]) until the detection of a malignancy progression (as defined in the protocol) or the occurrence of any fatality (whatever is first). If the observation of the patient was stopped (e.g. at study end) after a period and no event was observed, then this observation time t is called “censored.”
As used herein, the term “Time To Death” (also referred to as “TTD”) or “Overall Survival” (also referred to as “OS”) refers to a clinical endpoint frequently used in oncology trials (that includes but is not limited to clinical trials with reference to the present invention). The measurement for each patient equals the time elapsed from onset of the treatment of a patient in a trial (as defined in the protocol [i.e., see the examples infra]) until the occurrence of any fatality. If the observation of the patient is stopped (e.g. at study end) after a period t and the patient survived to this time, then this observation time t is called “censored.”
As used herein, the term “covariate” refers to certain variables or information relating to a patient. The clinical endpoints are frequently considered in regression models, where the endpoint represent the dependent variable and the biomarkers represent the main or target independent variables (regressors). If additional variables from the clinical data pool are considered these are denoted as (clinical) covariates. The term “clinical covariate” here is used to describe all clinical information about the patient, which are in general available at baseline. These clinical covariates comprise demographic information like sex, age etc., other anamnestic information, concomitant diseases, concomitant therapies, result of physical examinations, common laboratory parameters obtained, known properties of the target tumor, information quantifying the extent of malignant disease, clinical performance scores like ECOG or Karnofsky index, clinical disease staging, timing and result of pretreatments and disease history as well as all similar information, which may be associated with the clinical prognosis.
As used herein, the term “raw analysis” or “unadjusted analysis” refers to regression analyses, where over the considered biomarkers no additional clinical covariates were used in the regression model, neither as independent factors nor as stratifying covariate.
As used herein, the term “adjusted by covariates” refers to regression analyses, where over the considered biomarkers additional clinical covariates were used in the regression model, either as independent factors or as stratifying covariate.
As used herein, the term “univariate” refers to regression models or graphical approaches where as independent variable only one of the target biomarkers is part of the model. These univariate models can be considered with and without additional clinical covariates.
As used herein, the term “multivariate” refers to regression models or graphical approaches where as independent variables more than one of the target biomarkers are part of the model.
These multivariate models can be considered with and without additional clinical covariates.
“Nucleotides” are “nucleosides” that further include a phosphate group covalently linked to the sugar portion of the nucleoside. For those “nucleosides” that include a pentofuranosyl sugar, the phosphate group can be linked to either the 2′, 3′ or 5′ hydroxyl moiety of the sugar. A “nucleotide” is the “monomeric unit” of an “oligonucleotide”, more generally denoted herein as an “oligomeric compound”, or a “polynucleotide”, more generally denoted as a “polymeric compound”. Another general expression therefor is desoxyribonucleic acid (DNA) and ribonucleic acid (RNA). As used herein the term “polynucleotide” is synonymous with “nucleic acid.”
As used herein, the term “probe” refers to synthetically or biologically produced nucleic acids (DNA or RNA) which, by design or selection, contain specific nucleotide sequences that allow them to hybridize under defined predetermined stringencies specifically (i.e., preferentially) to “nucleic acids”. A “probe” can be identified as a “capture probe” meaning that it “captures” the nucleic acid so that it can be separated from undesirable materials which might obscure its detection. Once separation is accomplished, detection of the captured “target nucleic acid” can be achieved using a suitable procedure. “Capture probes” are often already attached to a solid phase. According to the present invention, the term hybridization under “stringent conditions” is given the same meaning as in Sambrook et al. (Molecular Cloning, A Laboratory Manual, Cold Spring Harbor Laboratory Press (1989), paragraph 1.101-1.104). Preferably, a “stringent hybridization” is the case when a hybridization signal is still detectable after washing for 1 h with 1×SSC and 0.1% SDS at 50° C., preferably at 55° C., more preferably at 62° C., and most preferably at 68° C., and more preferably for 1 hour with 0.2×SSC and 0.1% SDS at 50°., preferably at 55° C., more preferably at 62°, and most preferably at 68° C. The composition of the SSC buffer is described in Sambrook et al. (Molecular Cloning, A Laboratory Manual, Cold Spring Harbor Laboratory Press (1989)).
As used herein, a “transcribed polynucleotide” is a polynucleotide (e.g an RNA, a cDNA, or an analog of one of an RNA or cDNA) which is complementary to or homologous with all or a portion of a mature RNA made by transcription of a gene, such as the marker gene of the invention, and normal post-transcriptional processing (e.g. splicing), if any, of the transcript. The term “cDNA” is an abbreviation for complementary DNA, the single-stranded or double-stranded DNA copy of a mRNA. The term “mRNA” is an abbreviation for messenger RNA- the RNA that serves as a template for protein synthesis.
As used herein, the term “marker gene” or “biomarker gene” is meant to include a gene which is useful according to this invention for determining the progression of cancer in a patient, particularly in a breast cancer patient.
As used herein, the term “marker polynucleotide” or “biomarker polynucleotide” is meant to include a nucleotide transcript (hnRNA or mRNA) encoded by a marker gene according to the invention, or cDNA derived from the nucleotide transcript, or a segment of said transcript or cDNA.
As used herein, the term “marker protein,” “marker polypeptide,” “biomarker protein,” or “biomarker polypeptide” is meant to include a protein or polypeptide encoded by a marker gene according to the invention or to a fragment thereof.
As used herein, the term “marker” and “biomarker” are used interchangeably and refer to a marker gene, marker polynucleotide, or marker protein as defined above.
As used herein, the term “gene product” refers to a marker polynucleotide or marker protein encoded by a marker gene.
The expression of a marker gene “significantly” differs from the level of expression of the marker gene in a reference sample if the level of expression of the marker gene in a sample from the patient differs from the level in a sample from the reference subject by an amount greater than the standard error of the assay employed to assess expression, and preferably at least 10%, and more preferably 25%, 50%, 75%, 100%, 125%, 150%, 175%, 200%, 300%, 400%, 500% or 1,000% of that amount. Alternatively, expression of the marker gene in the patient can be considered “significantly” lower than the level of expression in a control subject if the level of expression in a sample from the patient is lower than the level in a sample from the control subject by an amount greater than the standard error of the assay employed to assess expression, and preferably at least 10%, and more preferably 25%, 50%, 75%, 100%, 125%, 150%, 175%, 200%, 300%, 400%, 500% or 1,000% that amount.
A marker polynucleotide or a marker protein “corresponds to” another marker polynucleotide or marker protein if it is related thereto, and in preferred embodiments is identical thereto.
The terms “level of expression” or “expression level” are used interchangeably and generally refer to the amount of a polynucleotide or an amino acid product or protein in a biological sample. “Expression” generally refers to the process by which gene encoded information is converted into the structures present and operating in the cell. Therefore, according to the invention “expression” of a gene may refer to transcription into a polynucleotide, translation into a protein or even posttranslational modification of the protein. Fragments of the transcribed polynucleotide, the translated protein or the postranslationally modified protein shall also be regarded as expressed whether they originate from a transcript generated by alternative splicing, a degraded transcript or from a posttranslational processing of the protein, e.g. by proteolysis. “Expressed genes” include those that are transcribed into a polynucleotide as mRNA and then translated into a protein; and also include expressed genes that are transcribed into RNA but not translated into a protein (for example, transfer and ribosomal RNAs).
The term “overexpression” or “increased expression” refers to an upward deviation in levels of expression as compared to the baseline expression level in a sample used as a control.
The term “underexpression” or “decreased expression” refers to a downward deviation in levels of expression as compared to the baseline expression level in a sample used as a control.
The term “amphiregulin” relates to a gene that encodes a protein and to the protein itself that is a member of the epidermal growth factor family. It is an autocrine growth factor as well as a mitogen for astrocytes, Schwann cells, and fibroblasts. It is related to epidermal growth factor (EGF) and transforming growth factor alpha (TGF-alpha). This protein interacts with the EGF/TGF-alpha receptor to promote the growth of normal epithelial cells and inhibits the growth of certain aggressive carcinoma cell lines. According to the invention, the amino acid sequence of amphiregulin is the amino acid sequence according to SEQ ID NO: 1. According to the invention, the nucleic acid sequence of the “amphiregulin” cDNA is the nucleic acid sequence according to SEQ ID NO: 5 which is accessible at GenBank with the accession number NM_—001657.
The term “transforming growth factor alpha” relates to a gene that encodes a protein and to the protein itself that is a member of the family of transforming growth factors (TGFs). These are biologically active polypeptides that reversibly confer the transformed phenotype on cultured cells. “Transforming growth factor-alpha” shows about 40% sequence homology with epidermal growth factor and competes with EGF for binding to the EGF receptor, stimulating its phosphorylation and producing a mitogenic response. According to the invention, the amino acid sequence of “Transforming growth factor-alpha” is the amino acid sequence according to SEQ ID NO: 3. According to the invention, the nucleic acid sequence of the “transforming growth factor-alpha” cDNA is the nucleic acid sequence according to SEQ ID NO: 7 which is accessible at GenBank with the accession number NM_—003236.
The term “epidermal growth factor” relates to a gene that encodes a protein and to the protein itself that is a member of the family of growth factors. “Epidermal growth factor (EGF)” has a profound effect on the differentiation of specific cells in vivo and is a potent mitogenic factor for a variety of cultured cells of both ectodermal and mesodermal origin. The EGF precursor is believed to exist as a membrane-bound molecule which is proteolytically cleaved to generate the 53-amino acid peptide hormone that stimulates cells to divide. According to the invention, the amino acid sequence of “Epidermal growth factor” is the amino acid sequence according to SEQ ID NO: 2. According to the invention, the nucleic acid sequence of the “Epidermal growth factor (EGF)” cDNA is the nucleic acid sequence according to SEQ ID NO: 6 which is accessible at GenBank with the accession number NM_—001963. The “Epidermal Growth Factor Receptor” abbreviated as EGFR, a 170-kD glycoprotein, is composed of an N-terminus extracellular domain, a hydrophobic transmembrane domain, and a C-terminus intracellular region containing the kinase domain. The mRNA has different variants translated into different receptor proteins. According to the invention, the amino acid sequence of the “Epidermal growth factor receptor” is the amino acid sequence according to SEQ ID NO: 11 (transcript variant 1; GenBank accession number NM_—005228), SEQ ID NO: 12 (transcript variant 2; GenBank accession number NM_—201282), SEQ ID NO: 13 (transcript variant 3; GenBank accession number NM_—201283), or SEQ ID NO: 14 (transcript variant 4; GenBank accession number NM_—201284). EGFR, encoded by the erbB1 gene, has been causally implicated in human malignancy. In particular, increased expression of EGFR has been observed in breast, bladder, lung, head, neck and stomach cancer as well as glioblastomas. EGFR ligandinduced dimerization activates the intrinsic RTK domain (an Src homology domain 1, SH1), resulting in autophosphorylation on six specific EGFR tyrosine residues in the noncatalytic tail of the cytoplasmic domain. The cellular effects of EGFR activation in a cancer cell include increased proliferation, promotion of cell motility, adhesion, invasion, angiogenesis, and enhanced cell survival by inhibition of apoptosis. Activated EGFR induces tumor cell proliferation through stimulation of the mitogen-activated protein kinase (MAPK) cascade.
The terms “human neu”, “c-erbB-2”, “erbB2”, “erbB-2”, “HER-2/neu”, “HER-2” and “HER2” are used interchangeably herein. These terms relate to a gene that encodes a protein and to the protein itself that is a member of the family of the epidermal growth factor (EGF) receptor family of receptor tyrosine kinases. This protein has no ligand binding domain of its own and therefore cannot bind growth factors. However, it does bind tightly to other ligand-bound EGF receptor family members to form a heterodimer, stabilizing ligand binding and enhancing kinase-mediated activation of downstream signalling pathways, such as those involving mitogen-activated protein kinase and phosphatidylinositol-3 kinase. Allelic variations at amino acid positions 654 and 655 of isoform a (positions 624 and 625 of isoform b) have been reported, with the most common allele, Ile654/Ile655 being preferred according to the invention. Amplification and/or overexpression of this gene has been reported in numerous cancers, including breast and ovarian tumors. Alternative splicing results in several additional transcript variants, some encoding different isoforms and others that have not been fully characterized. According to the invention, the amino acid sequence of HER2 is the amino acid sequence according to SEQ ID NO: 4. According to the invention, the nucleic acid sequence of the “HER2” cDNA is the nucleic acid sequence according to SEQ ID NO: 8 which is accessible at GenBank with the accession number NM_—004448.2.
The “extracellular domain of HER2” or “shed extracellular domain of HER2” or “HER2-ECD” is a glycoprotein of between 97 and 115 kDa which corresponds substantially to the extracellular domain of the human HER2 gene product. It can be referred to as p105 (Zabrecky, J. R. et al., J. Biol. Chem. 266 (1991) 1716-1720; U.S. Pat. No. 5,401,638; U.S. Pat. No. 5,604,107). The quantitation and detection of the extracellular domain of HER2 is described in U.S. Pat. No. 5,401,638 and U.S. Pat. No. 5,604,107.
The term “HER3” stands for another member of the epidermal growth factor receptor (EGFR) family of receptor tyrosine kinases. This membrane-bound protein has not an active kinase domain. The protein can bind ligands but not transmit a signal into the cell. It forms heterodimers with other EGF receptor family members which do have kinase activity which leads to cell proliferation or differentiation. Amplification of this gene and/or overexpression of its protein is found in numerous cancers. According to the invention, the amino acid sequence of the “HER3” cDNA is the amino acid sequence according to SEQ ID NO: 9 which is accessible at GenBank from the translation of the nucleic acid sequence of HER3 with the accession number NM_—001005915. According to the invention, the nucleic acid sequence of the “HER3” cDNA is the nucleic acid sequence according to SEQ ID NO: 10 which is accessible at GenBank with the accession number NM_—001005915.
The term “antibody” herein is used in the broadest sense and specifically covers intact monoclonal antibodies, polyclonal antibodies, and multispecific antibodies (e.g., bispecific antibodies) formed from at least two intact antibodies, and antibody fragments, so long as they exhibit the desired biological activity of an antibody.
The term “monoclonal antibody” as used herein refers to an antibody obtained from a population of substantially homogeneous antibodies, i.e., the individual antibodies comprising the population are identical except for possible naturally occurring mutations that may be present in minor amounts. Monoclonal antibodies are highly specific, being directed against a single antigenic site. Furthermore, in contrast to polyclonal antibody preparations which include different antibodies directed against different determinants (epitopes), each monoclonal antibody is directed against a single determinant on the antigen. In addition to their specificity, the monoclonal antibodies are advantageous in that they may be synthesized uncontaminated by other antibodies. The modifier “monoclonal” indicates the character of the antibody as being obtained from a substantially homogeneous population of antibodies, and is not to be construed as requiring production of the antibody by any particular method. For example, the monoclonal antibodies to be used in accordance with the present invention may be made by the hybridoma method first described by Kohler, G. et al., Nature 256 (1975) 495-497, or may be made by recombinant DNA methods (see, e.g., U.S. Pat. No. 4,816,567). “Antibody fragments” comprise a portion of an intact antibody.
An antibody “which binds” an antigen of interest according to the invention is one capable of binding that antigen with sufficient affinity such that the antibody is useful in detecting the presence of the antigen. One antibody according to the invention binds human HER2 and does not (significantly) cross-react with other proteins. In such embodiments, the extent of binding of the antibody to other proteins will be less than 10% as determined by fluorescence activated cell sorting (FACS) analysis or radioimmunoprecipitation (RIA).
Dimerization—the pairing of receptors—is essential to the signaling activity of all HER receptors. According to the invention, the term “HER dimerization inhibitor” or preferably “HER2 heterodimerization inhibitor” refers to a therapeutic agent that binds to HER2 and inhibits HER2 heterodimerization. These are preferably antibodies, preferably monoclonal antibodies, more preferably humanized antibodies that bind to HER2 and inhibit HER2 heterodimerization. Examples of antibodies that bind HER2 include 4D5, 7C2, 7F3 or 2C4 as well as humanized variants thereof, including huMAb4D5-1, huMAb4D5-2, huMAb4D5-3, huMAb4D5-4, huMAb4D5-5, huMAb4D5-6, huMAb4D5-7 and huMAb4D5-8 as described in Table 3 of U.S. Pat. No. 5,821,337; and humanized 2C4 mutant numbers 560, 561, 562, 568, 569, 570, 571, 574, or 56869 as described in WO 01/00245. 7C2 and 7F3 and humanized variants thereof are described in WO 98/17797. The term “HER dimerization inhibitor” or “HER2 heterodimerization inhibitor” shall not apply to Trastuzumab monoclonal antibodies commercially available as “Herceptin®” as the mechanism of action is different and as Trastuzumab does not inhibit HER dimerization.
Preferred throughout the application is the “antibody 2C4”, in particular the humanized variant thereof (WO 01/00245; produced by the hybridoma cell line deposited with the American Type Culture Collection, Manassass, Va., USA under ATCC HB-12697), which binds to a region in the extracellular domain of HER2 (e.g., any one or more residues in the region from about residue 22 to about residue 584 of HER2, inclusive). The “epitope 2C4” is the region in the extracellular domain of ErbB2 to which the antibody 2C4 binds. The expression “monoclonal antibody 2C4” refers to an antibody that has antigen binding residues of, or derived from, the murine 2C4 antibody of the Examples in WO 01/00245. For example, the monoclonal antibody 2C4 may be murine monoclonal antibody 2C4 or a variant thereof, such as humanized antibody 2C4, possessing antigen binding amino acid residues of murine monoclonal antibody 2C4. Examples of humanized 2C4 antibodies are provided in Example 3 of WO 01/00245. Unless indicated otherwise, the expression “rhuMAb 2C4” when used herein refers to an antibody comprising the variable light (VL) and variable heavy (VH) sequences of SEQ ID Nos. 3 and 4 of WO 01/00245, respectively, fused to human light and heavy IgG1 (non-A allotype) constant region sequences optionally expressed by a Chinese Hamster Ovary (CHO) cell. Preferred embodiments of WO 01/00245 are preferred herein as well. The humanized antibody 2C4 is also called Pertuzumab.
A “kit” is any manufacture (e.g a package or container) comprising at least one reagent, e.g a probe, for specifically detecting a marker gene or protein of the invention. The manufacture is preferably promoted, distributed, or sold as a unit for performing the methods of the present invention.
The verbs “determine” and “assess” shall have the same meaning and are used interchangeably throughout the application.
Conventional techniques of molecular biology and nucleic acid chemistry, which are within the skill of the art, are explained in the literature. See, for example, Sambrook, J. et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., 1989; Gait, M. J. (ed.), Oligonucleotide synthesis—a practical approach, IRL Press Limited, 1984; Hames, B. D. and Higgins, S. J. (eds.), Nucleic acid hybridisation—a practical approach, IRL Press Limited, 1985; and a series, Methods in Enzymology, Academic Press, Inc., all of which are incorporated herein by reference. All patents, patent applications, and publications mentioned herein, both supra and infra, are hereby incorporated by reference in their entirety.
As used herein, the general form of a prediction rule consists in the specification of a function of one or multiple biomarkers potentially including clinical covariates to predict response or non-response, or more generally, predict benefit or lack of benefit in terms of suitably defined clinical endpoints.
The simplest form of a prediction rule consists of an univariate model without covariates, where the prediction is determined by means of a cutoff or threshold. This can be phrased in terms of the Heaviside function for a specific cutoff c and a biomarker measurement x, where the binary prediction A or B is to be made, then
If H(x−c)=0 then predict A.
If H(x−c)=1 then predict B.
This is the simplest way of using univariate biomarker measurements in prediction rules. If such a simple rule is sufficient, it allows for a simple identification of the direction of the effect, i.e. whether high or low expression levels are beneficial for the patient.
The situation can be more complicated if clinical covariates need to be considered and/or if multiple biomarkers are used in multivariate prediction rules. In order to illustrate the issues here are two hypothetical examples:

Covariate Adjustment (Hypothetical Example):

For a biomarker X it is found in a clinical trial population that high expression levels are associated with a worse prognosis (univariate analysis). A closer analysis shows that there are two tumor types in the population, one of which possess a worse prognosis than the other one and at the same time the biomarker expression for this tumor group is generally higher. An adjusted covariate analysis reveals that for each of the tumor types the relation of clinical benefit and prognosis is reversed, i.e. within the tumor types, lower expression levels are associated with better prognosis. The overall opposite effect was masked by the covariate tumor type—and the covariate adjusted analysis as part of the prediction rule reversed the direction.

Multivariate Prediction (Hypothetical Example):

For a biomarker X it is found in a clinical trial population that high expression levels are slightly associated with a worse prognosis (univariate analysis). For a second biomarker Y a similar observation was made by univariate analysis. The combination of X and Y revealed that a good prognosis is seen if both biomarkers are low. This makes the rule to predict benefit if both biomarkers are below some cutoffs (AND- connection of a Heaviside prediction function). For the combination rule there is no longer a simple rule phraseable in an univariate sense. E.g. having low expression levels in X will not automatically predict a better prognosis.
These simple examples show that prediction rules with and without covariates cannot be judged on the univariate level of each biomarker. The combination of multiple biomarkers plus a potential adjustment by covariates does not allow to assign simple relationships towards single biomarkers.
In one embodiment of the invention, a method of predicting the response to a treatment with a HER inhibitor, preferably a HER dimerization inhibitor, in a patient comprises the steps of:

- (a) determining the expression level or amount of one or more biomarkers in a biological sample from a patient wherein the biomarker or biomarkers are selected from the group consisting of:
  - (1) transforming growth factor alpha;
  - (2) HER2;
  - (3) amphiregulin; and
  - (4) epidermal growth factor;
- (b) determining whether the expression level or amount assessed in step (a) is above or below a certain quantity that is associated with an increased or decreased clinical benefit to a patient; and
- (c) predicting the response to the treatment with the HER inhibitor in the patient by evaluating the results of step (b).

In a more particular embodiment of the above method, the expression level of the transforming growth factor alpha biomarker is determined in combination with one or more biomarkers selected from the group consisting of epidermal growth factor, amphiregulin, and HER2. In another more particular embodiment of the above method, the expression level of the HER2 biomarker is determined in combination with one or more biomarkers selected from the group consisting of epidermal growth factor, transforming growth factor alpha, and amphiregulin. In another more particular embodiment of the above method, the expression level of the epidermal growth factor biomarker is determined in combination with one or more biomarkers selected from the group consisting of amphiregulin, transforming growth factor alpha, and HER2. In another more particular embodiment of the above method, an amphiregulin biomarker is assessed in combination with one or more biomarkers selected from the group consisting of epidermal growth factor, transforming growth factor alpha, and HER2.
The “quantity that is associated with an increased or decreased clinical benefit to a patient” of the above method is preferably a value expressed in mass/volume for blood serum or blood plasma or mass/mass for tumor tissue. It can be measured by methods known to the expert skilled in the art and also disclosed by this invention. If the expression level or amount determined in step (a) is above or below a certain quantity or value, the response to the treatment can be determined.
With respect to the quantity in blood serum for the transforming growth factor alpha marker protein, a range between 2.0-10.0 pg/ml, preferably a range between 2.0-5.0 pg/ml, and more preferably about 3.5 pg/ml may be favorable for progression free survival and overall survival when treatment with a HER inhibitor is considered. See FIG. 7. Thus, in a preferred embodiment, the quantity of transforming growth factor alpha marker protein in the blood serum of a patient is within one of the foregoing ranges for predicting a good response to treatment with a HER inhibitor in the patient.
With respect to the quantity in blood serum for the HER2 marker protein (preferably the soluble HER2 extracellular domain (HER2-ECD)), a range between 12-22 ng/ml, preferably about 18 ng/ml, may be favorable for progression free survival and overall survival when treatment with a HER inhibitor is considered. See FIG. 7. Thus, in a preferred embodiment, the quantity of HER2 marker protein in the blood serum of a patient is within the foregoing range for predicting a good response to treatment with a HER inhibitor in the patient.
With respect to the quantity in blood serum for the epidermal growth factor marker protein, a range between 100-250 pg/ml, preferably about 150 pg/ml, may be favorable for progression free survival and overall survival when treatment with a HER inhibitor is considered. See FIG. 7. Thus, in a preferred embodiment, the quantity of epidermal growth factor marker protein in the blood serum of a patient is within the foregoing range for predicting a good response to treatment with a HER inhibitor in the patient.
With respect to the quantity in blood serum for the amphiregulin marker protein, a range between 6-15 pg/ml, preferably about 12 pg/ml, may be favorable for progression free survival and overall survival when treatment with a HER inhibitor is considered. See FIG. 7. Thus, in a preferred embodiment, the quantity of amphiregulin marker protein in the blood serum of a patient is within the foregoing range for predicting a good response to treatment with a HER inhibitor in the patient.
Since the marker genes, in particular in serum, may be used in multiple-marker prediction models potentially including other clinical covariates, the direction of a beneficial effect of a single marker gene within such models cannot be determined in a simple way, and may contradict the direction found in univariate analyses, i.e. the situation as described for the single marker gene.
More preferably, in the method according to the invention, the quantity or value (below or above which is associated with an increased or decreased clinical benefit) is determined by:

- (1) determining the expression level or amount of a biomarker or combination of biomarkers in a plurality of biological samples from patients before treatment with the HER inhibitor,
- (2) treating the patients with the HER inhibitor,
- (3) determining the clinical benefit of each patient; and
- (4) correlating the clinical benefit of the patients treated with the HER inhibitor to the expression level or amount of the biomarker or combination of biomarkers.

The “quantity” is preferably a value expressed in mass/volume for blood serum or blood plasma or mass/mass for tumor tissue.
The present invention also considers mutants or variants of the marker genes according to the present invention and used in the methods according to the invention. In those mutants or variants the native sequence of the marker gene is changed by substitutions, deletions or insertions. “Native sequence” refers to an amino acid or nucleic acid sequence which is identical to a wild-type or native form of a marker gene or protein.
The present invention also considers mutants or variants of the proteins according to the present invention and used in the methods according to the invention. “Mutant amino acid sequence,” “mutant protein” or “mutant polypeptide” refers to a polypeptide having an amino acid sequence which varies from a native sequence or is encoded by a nucleotide sequence intentionally made variant from a native sequence. “Mutant protein,” “variant protein” or “mutein” means a protein comprising a mutant amino acid sequence and includes polypeptides which differ from the amino acid sequence of the native protein according to the invention due to amino acid deletions, substitutions, or both.
The present invention also considers a method of predicting the response to a treatment with a combination of a HER inhibitor and another substance or agent as a chemotherapeutic agent or a therapeutic antibody used for treating cancer. The chemotherapeutic agent may be e.g. gemcitabine (Gemzar®; chemical name: 2′,2′-difluorodeoxycytidine (dFdC)), carboplatin (diammine-(cyclobutane-1,1-dicarboxylato (2-)-O,O′)-platinum), or paclitaxel (Taxol®, chemical name: β-(benzoylamino)-α-hydroxy-,6,12b-bis(acetyloxy)-12-(benzoyloxy)-2a,3,4,4a,5,6,9,10,11,12,12a,12b-dodecahydro-4,11-dihydroxy-4a,8,13,13-tetramethyl-5-oxo-7,11-methano-1H-cyclodeca(3,4)benz(1,2-b)oxet-9-yl ester,(2aR-(2a-α,4-β,4a-β,6-β, 9-α(α-R*,β-S*),11-α,12-α,12a-α, 2b-α))-benzenepropanoic acid); or transtuzumab; or erlotinib.
In a preferred embodiment of the invention, the biological sample is blood serum, blood plasma or tumor tissue. Tumor tissue may be formalin-fixed paraffin embedded tumor tissue or fresh frozen tumor tissue.
In another preferred embodiment of the invention, the HER dimerization inhibitor inhibits heterodimerization of HER2 with EGFR or HER3, or HER4. Preferably, the HER dimerization inhibitor is an antibody, preferably the antibody 2C4. Preferred throughout the application is the “antibody 2C4”, in particular the humanized variant thereof (WO 01/00245; produced by the hybridoma cell line deposited with the American Type Culture Collection, Manassass, Va., USA under ATCC HB-12697), which binds to a region in the extracellular domain of HER2 (e.g., any one or more residues in the region from about residue 22 to about residue 584 of HER2, inclusive). Examples of humanized 2C4 antibodies are provided in Example 3 of WO 01/00245. The humanized antibody 2C4 is also called Pertuzumab.
In still another preferred embodiment of the invention, the patient is a cancer patient, preferably a breast cancer, ovarian cancer, lung cancer or prostate cancer patient. The breast cancer patient is preferably a metastatic breast cancer patient or a HER2 low expressing breast or metastatic breast cancer patient, or a HER2 high expressing breast or metastatic breast cancer patient. The ovarian cancer patient is preferably a metastatic ovarian cancer patient. The lung cancer patient is preferably a non-small cell lung cancer (NSCLC) patient.
It is preferred that two, three or all four marker genes, marker polynucleotides or marker proteins are used in combination, i.e. used in all disclosed embodiments of the invention or methods, uses or kits according to the invention. The following are preferred combinations of biomarkers in which the level of expression or amounts are determined in accordance with the invention:
In one particular embodiment, a transforming growth factor alpha biomarker is assessed in combination with one or more biomarkers selected from the group consisting of epidermal growth factor, amphiregulin, and HER2. In another particular embodiment, a HER2 biomarker is assessed in combination with one or more biomarkers selected from the group consisting of epidermal growth factor, transforming growth factor alpha, and amphiregulin.
In another particular embodiment, a epidermal growth factor biomarker is assessed in combination with one or more biomarkers selected from the group consisting of amphiregulin, transforming growth factor alpha, and HER2. In another particular embodiment, an amphiregulin biomarker is assessed in combination with one or more biomarkers selected from the group consisting of epidermal growth factor, transforming growth factor alpha, and HER2.
In a particularly preferred embodiment of the invention, the combination of biomarkers consists of:

- the transforming growth factor alpha and the HER2 biomarkers, or
- the transforming growth factor alpha and the EGF biomarkers, or
- the amphiregulin, the epidermal growth factor, the transforming growth factor alpha and the HER2 biomarkers,

In a preferred embodiment of the invention, the level of expression of the marker gene or the combination of marker genes in the sample is assessed by detecting the level of expression of a marker protein or a fragment thereof or a combination of marker proteins or fragments thereof encoded by the marker gene or the combination of marker genes. Preferably, the level of expression of the marker protein or the fragment thereof or the combination of marker proteins or the fragments thereof is detected using a reagent which specifically binds with the marker protein or the fragment thereof or the combination of marker proteins or the fragments thereof. Preferably, the reagent is selected from the group consisting of an antibody, a fragment of an antibody, and an antibody derivative.
There are many different types of immunoassays which may be used in the method of the present invention, e.g. enzyme linked immunoabsorbent assay (ELISA), fluorescent immunosorbent assay (FIA), chemical linked immunosorbent assay (CLIA), radioimmuno assay (RIA), and immunoblotting. For a review of the different immunoassays which may be used, see: Lottspeich and Zorbas (eds.), Bioanalytik, 1^stedition 1998, Spektrum Akademischer Verlag, Heidelberg, Berlin, Germany. Therefore, in yet another preferred embodiment of the invention, the level of expression is determined using a method selected from the group consisting of proteomics, flow cytometry, immunocytochemistry, immunohistochemistry, enzyme-linked immunosorbent assay, multi-channel enzyme-linked immunosorbent assay, and variations of these methods. Therefore more preferably, the level of expression is determined using a method selected from the group consisting of proteomics, flow cytometry, immunocytochemistry, immunohistochemistry, enzyme-linked immunosorbent assay, multi-channel enzyme-linked immunosorbent assay, and variations of these methods.
In another preferred embodiment of the invention, the fragment of the marker protein is the extracellular domain of the HER2 marker protein (HER2-ECD). Preferably, the extracellular domain of the HER2 marker protein has a molecular mass of approximately 105,000 Dalton. “Dalton” stands for a mass unit that is equal to the weight of a hydrogen atom, or 1.657×10⁻²⁴grams.
In another preferred embodiment of the invention

- the amino acid sequence of the amphiregulin marker protein is the amino acid sequence SEQ ID NO: 1,
- the amino acid sequence of the epidermal growth factor marker protein is the amino acid sequence SEQ ID NO: 2,
- the amino acid sequence of the transforming growth factor alpha marker protein is the amino acid sequence SEQ ID NO: 3, or
- the amino acid sequence of the HER2 marker protein is the amino acid sequence SEQ ID NO: 4.

In another preferred embodiment of the invention, the quantity in blood serum for

- the transforming growth factor alpha marker protein is between 2.0 to 10.0 pg/ml, preferably about 3.5 pg/ml,
- the epidermal growth factor marker protein is between 100 to 250 pg/ml, preferably about 150 pg/ml, or pg/ml.
- the amphiregulin marker protein is between 6 to 15 pg/ml, preferably about 12 the HER2 marker protein is between 12 to 22 ng/ml, preferably about 18 ng/ml.

In still another preferred embodiment of the invention, the “quantity” in blood serum for the extracellular domain of the HER2 marker protein is between 12 to 22 ng/ml, preferably about 18 ng/ml.
In yet another preferred embodiment of the invention, the level of expression of the marker gene or the combination of marker genes in the biological sample is assessed by detecting the level of expression of a transcribed marker polynucleotide encoded by the marker gene or a fragment of the transcribed marker polynucleotide or of transcribed marker polynucleotides encoded by the combination of marker genes or fragments of the transcribed marker polynucleotide. Preferably, the transcribed marker polynucleotide is a cDNA, mRNA or hnRNA or wherein the transcribed marker polynucleotides are cDNA, mRNA or hnRNA.
Preferably, the step of detecting further comprises amplifying the transcribed polynucleotide. The amplification is performed preferably with the polymerase chain reaction which specifically amplifies nucleic acids to detectable amounts. Other possible amplification reactions are the Ligase Chain Reaction (LCR; Wu D. Y. and Wallace R. B., Genomics 4 (1989) 560-569; and Barany F., Proc. Natl. Acad. Sci. USA 88 (1991)189-193); Polymerase Ligase Chain Reaction (Barany F., PCR Methods and Applic. 1 (1991) 5-16); Gap-LCR (WO 90/01069); Repair Chain Reaction (EP 0439182 A2), 3SR (Kwoh, D. Y. et al., Proc. Natl. Acad. Sci. USA 86 (1989) 1173-1177; Guatelli, J. C. et al., Proc. Natl. Acad. Sci. USA 87 (1990) 1874-1878; WO 92/08808), and NASBA (U.S. Pat. No. 5,130,238). Further, there are strand displacement amplification (SDA), transcription mediated amplification (TMA), and Q(3-amplification (for a review see e.g. Whelen, A. C. and Persing, D. H., Annu. Rev. Microbiol. 50 (1996) 349-373; Abramson, R. D. and Myers T. W., Curr. Opin. Biotechnol. 4 (1993) 41-47). More preferably, the step of detecting is using the method of quantitative reverse transcriptase polymerase chain reaction.
Other suitable polynucleotide detection methods are known to the expert in the field and are described in standard textbooks as Sambrook J. et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., 1989; and Ausubel, F. et al., Current Protocols in Molecular Biology, 1987, J. Wiley and Sons, NY. There may be also further purification steps before the polynucleotide detection step is carried out as e.g. a precipitation step. The detection methods may include but are not limited to the binding or intercalating of specific dyes as ethidiumbromide which intercalates into the double-stranded polynucleotides and changes their fluorescence thereafter. The purified polynucleotide may also be separated by electrophoretic methods optionally after a restriction digest and visualized thereafter. There are also probe-based assays which exploit the oligonucleotide hybridisation to specific sequences and subsequent detection of the hybrid. It is also possible to sequence the DNA after further steps known to the expert in the field. The preferred template-dependent DNA polymerase is Taq polymerase.
In yet another preferred embodiment of the invention, the level of expression of the marker gene is assessed by detecting the presence of the transcribed marker polynucleotide or the fragment thereof in a sample with a probe which anneals with the transcribed marker polynucleotide or the fragment thereof under stringent hybridization conditions or the level of expression of the combination of the marker genes in the samples is assessed by detecting the presence of transcribed marker polynucleotides or the fragments thereof in a sample with probes which anneal with the transcribed marker polynucleotides or the fragments thereof under stringent hybridization conditions. This method may be performed in a homogeneous assay system. An example for a “homogeneous” assay system is the TagMan® system that has been detailed in U.S. Pat. No. 5,210,015, U.S. Pat. No. 5,804,375 and U.S. Pat. No. 5,487,972. Briefly, the method is based on a double-labelled probe and the 5′-3′ exonuclease activity of Taq DNA polymerase. The probe is complementary to the target sequence to be amplified by the PCR process and is located between the two PCR primers during each polymerisation cycle step. The probe has two fluorescent labels attached to it. One is a reporter dye, such as 6-carboxyfluorescein (FAM), which has its emission spectra quenched by energy transfer due to the spatial proximity of a second fluorescent dye, 6-carboxy-tetramethyl-rhodamine (TAMRA). In the course of each amplification cycle, the Taq DNA polymerase in the process of elongating a primed DNA strand displaces and degrades the annealed probe, the latter due to the intrinsic 5′-3′ exonuclease activity of the polymerase. The mechanism also frees the reporter dye from the quenching activity of TAMRA. As a consequence, the fluorescent activity increases with an increase in cleavage of the probe, which is proportional to the amount of PCR product formed. Accordingly, an amplified target sequence is measured by detecting the intensity of released fluorescence label. Another example for “homogeneous” assay systems are provided by the formats used in the LightCycler® instrument (see e.g. U.S. Pat. No. 6,174,670), some of them sometimes called “kissing probe” formats. Again, the principle is based on two interacting dyes which, however, are characterized in that the emission wavelength of a donor-dye excites an acceptor-dye by fluorescence resonance energy transfer. The COBAS® AmpliPrep instrument (Roche Diagnostics GmbH, D-68305 Mannheim, Germany) was recently introduced to expand automation by isolating target sequences using biotinylated sequence-specific capture probes along with streptavidin-coated magnetic particles (Jungkind, D., J. Clin. Virol. 20 (2001) 1-6; Stelzl, E. et al., J. Clin. Microbiol. 40 (2002) 1447-1450). It has lately been joined by an additional versatile tool, the Total Nucleic Acid Isolation (TNAI) Kit (Roche Diagnostics). This laboratory-use reagent allows the generic, not sequence-specific isolation of all nucleic acids from plasma and serum on the COBAS® AmpliPrep instrument based essentially on the method developed by Boom, R. et al., J. Clin. Microbiol. 28 (1990) 495-503.
In another preferred embodiment of the invention, the nucleic acid sequence of the amphiregulin marker polynucleotide is the nucleic acid sequence SEQ ID NO: 5, the nucleic acid sequence of the epidermal growth factor marker polynucleotide is the nucleic acid sequence SEQ ID NO: 6, the nucleic acid sequence of the transforming growth factor alpha marker polynucleotide is the nucleic acid sequence SEQ ID NO: 7, or the nucleic acid sequence of the HER2 marker polynucleotide is the nucleic acid sequence SEQ ID NO: 8.
In another embodiment of the invention, a probe that hybridizes with the epidermal growth factor, transforming growth factor alpha or HER2 marker polynucleotide under stringent conditions or an antibody that binds to the epidermal growth factor, transforming growth factor alpha or HER2 marker protein is used for predicting the response to treatment with a HER inhibitor in a patient or a probe that hybridizes with the amphiregulin, epidermal growth factor, transforming growth factor alpha or HER2 marker polynucleotide under stringent conditions or an antibody that binds to the amphiregulin, epidermal growth factor, transforming growth factor alpha or HER2 marker protein is used for selecting a composition for inhibiting the progression of disease in a patient. The disease is preferably cancer and the patient is preferably a cancer patient as disclosed above.
In another embodiment of the invention, a kit comprising a probe that anneals with the amphiregulin, epidermal growth factor, transforming growth factor alpha or HER2 marker polynucleotide under stringent conditions or an antibody that binds to the amphiregulin, epidermal growth factor, transforming growth factor alpha or HER2 marker protein is provided. Such kits known in the art further comprise plastics ware which can be used during the amplification procedure as e.g. microtitre plates in the 96 or 384 well format or just ordinary reaction tubes manufactured e.g. by Eppendorf, Hamburg, Germany and all other reagents for carrying out the method according to the invention, preferably an immunoassay, e.g. enzyme linked immunoabsorbent assay (ELISA), fluorescent immunosorbent assay (FIA), chemical linked immunosorbent assay (CLIA), radioimmuno assay (RIA), and immunoblotting. For a review of the different immunoassays and reagents which may be used, see: Lottspeich and Zorbas (eds.), Bioanalytik, 1^stedition, 1998, Spektrum Akademischer Verlag, Heidelberg, Berlin, Germany. Preferably combinations of the probes or antibodies to the various marker polynucleotides or marker proteins are provided in the form of kit as the preferred combinations of the marker polynucleotides or marker proteins as disclosed above.
In another embodiment of the invention, a method of selecting a composition for inhibiting the progression of disease in a patient is provided, the method comprising:

- (a) separately exposing aliquots of a biological sample from a cancer patient in the presence of a plurality of test compositions;
- (b) comparing the level of expression of one or more biomarkers selected from the group consisting of amphiregulin, epidermal growth factor, transforming growth factor alpha and HER2 in the aliquots of the biological sample contacted with the test compositions and the level of expression of such biomarkers in an aliquot of the biological sample not contacted with the test compositions; and
- (c) selecting one of the test compositions which alters the level of expression of the biomarker or biomarkers in the aliquot containing that test composition relative to the aliquot not contacted with the test composition wherein an at least 10% difference between the level of expression of the biomarker or biomarkers in the aliquot of the biological sample contacted with the test composition and the level of expression of the corresponding biomarker or biomarkers in the aliquot of the biological sample not contacted with the test composition is an indication for the selection of the test composition. The disease is preferably cancer and the patient is preferably a cancer patient as disclosed above.

In another embodiment of the invention, a method of selecting a composition for inhibiting the progression of disease in a patient is provided, the method comprising:

- (a) separately exposing aliquots of a biological sample from a cancer patient in the presence of a plurality of test compositions;
- (b) comparing the level of expression of one or more biomarkers selected from the group consisting of the amphiregulin, epidermal growth factor, transforming growth factor alpha and HER2 in the aliquots of the biological sample contacted with the test compositions and the level of expression of such biomarkers in an aliquot of the biological sample not contacted with the test compositions; and
- (c) selecting one of the test compositions which alters the level of expression of the biomarker or biomarkers in the aliquot containing that test composition relative to the aliquot not contacted with the test composition wherein an at least 10% difference between the level of expression of the biomarker or biomarkers in the aliquot of the biological sample contacted with the test composition and the level of expression of the corresponding biomarker or biomarkers in the aliquot of the biological sample not contacted with the test composition is an indication for the selection of the test composition. The disease is preferably cancer and the patient is preferably a cancer patient as disclosed above.

The expression of a marker gene “significantly” differs from the level of expression of the marker gene in a reference sample if the level of expression of the marker gene in a sample from the patient differs from the level in a sample from the reference subject by an amount greater than the standard error of the assay employed to assess expression, and preferably at least 10%, and more preferably 25%, 50%, 75%, 100%, 125%, 150%, 175%, 200%, 300%, 400%, 500% or 1,000% of that amount. Alternatively, expression of the marker gene in the patient can be considered “significantly” lower than the level of expression in a reference subject if the level of expression in a sample from the patient is lower than the level in a sample from the reference subject by an amount greater than the standard error of the assay employed to assess expression, and preferably at least 10%, and more preferably 25%, 50%, 75%, 100%, 125%, 150%, 175%, 200%, 300%, 400%, 500% or 1,000% that amount. The difference of the level of expression be up to 10,000 or 50,000%. The difference of the level of expression is preferably between 10% to 10,000%, more preferably 25% to 10,000%, 50% to 10,000%, 100% to 10,000%, even more preferably 25% to 5,000%, 50% to 5,000%, 100% to 5,000%.
In another embodiment of the invention, a method of identifying a candidate agent is provided said method comprising:

- (a) contacting an aliquot of a biological sample from a cancer patient with the candidate agent and determining the level of expression of one or more biomarkers selected from the group consisting of amphiregulin, epidermal growth factor, transforming growth factor alpha and HER2 in the aliquot;
- (b) determining the level of expression of a corresponding biomarker or biomarkers in an aliquot of the biological sample not contacted with the candidate agent;
- (c) observing the effect of the candidate agent by comparing the level of expression of the biomarker or biomarkers in the aliquot of the biological sample contacted with the candidate agent and the level of expression of the corresponding biomarker or biomarkers in the aliquot of the biological sample not contacted with the candidate agent; and
- (d) identifying said agent from said observed effect, wherein an at least 10% difference between the level of expression of the biomarker gene or combination of biomarker genes in the aliquot of the biological sample contacted with the candidate agent and the level of expression of the corresponding biomarker gene or combination of biomarker genes in the aliquot of the biological sample not contacted with the candidate agent is an indication of an effect of the candidate agent.

In still another embodiment of the invention, a method of identifying a candidate agent is provided said method comprising:

- (a) contacting an aliquot of a biological sample from a cancer patient with the candidate agent and determining the level of expression in the aliquot of:
  - (1) a biomarker or a combination of biomarkers selected from the group consisting of epidermal growth factor, transforming growth factor alpha and HER2 or;
  - (2) a combination of biomarkers comprising amphiregulin and one or more biomarkers selected from the group consisting of an epidermal growth factor, a transforming growth factor alpha, and HER2,
- (b) determining the level of expression of a corresponding biomarker or biomarkers in an aliquot of the biological sample not contacted with the candidate agent,
- (c) observing the effect of the candidate agent by comparing the level of expression of the biomarker or biomarkers in the aliquot of the biological sample contacted with the candidate agent and the level of expression of the corresponding biomarker or biomarkers in the aliquot of the biological sample not contacted with the candidate agent,
- (d) identifying said agent from said observed effect, wherein an at least 10% difference between the level of expression of the biomarker or biomarkers in the aliquot of the biological sample contacted with the candidate agent and the level of expression of the corresponding biomarker or biomarkers in the aliquot of the biological sample not contacted with the candidate agent is an indication of an effect of the candidate agent.

Preferably, the candidate agent is a candidate inhibitory agent. Preferably, said candidate agent is a candidate enhancing agent.
In another embodiment of the invention, a candidate agent derived by the method according to the invention is provided.
In another embodiment of the invention, a pharmaceutical preparation comprising an agent according to the invention is provided.
In yet another embodiment of the invention, an agent according to the invention is used for the preparation of a composition for the treatment of cancer. Preferred forms of cancer are disclosed above.
In another preferred embodiment of the invention, a method of producing a drug comprising the steps of the method according to the invention and

- (i) synthesizing the candidate agent identified in step (c) above or an analog or derivative thereof in an amount sufficient to provide said drug in a therapeutically effective amount to a subject; and/or
- (ii) combining the drug candidate the candidate agent identified in step (c) above or an analog or derivative thereof with a pharmaceutically acceptable carrier.

In another embodiment of the invention, a marker protein or a marker polynucleotide selected from the group consisting of a amphiregulin, epidermal growth factor, transforming growth factor alpha and HER2 marker protein or marker polynucleotide is used for identifying a candidate agent or for selecting a composition for inhibiting the progression of a disease in a patient. The disease is preferably cancer and the patient is preferably a cancer patient as disclosed above.
In another embodiment of the invention, a HER inhibitor is used for the manufacture of a pharmaceutical composition for treating a human cancer patient characterized in that said treating or treatment includes assessing in a biological sample from the patient

- (a) a marker gene or a combination of marker genes selected from the group consisting of an epidermal growth factor, a transforming growth factor alpha and a HER2 marker gene or;
- (b) a combination of marker genes comprising an amphiregulin marker gene and a marker gene selected from the group consisting of an epidermal growth factor, a transforming growth factor alpha and a HER2 marker gene.

The manufacture of a pharmaceutical composition for treating a human cancer patient and particularly the formulation is described in WO 01/00245, incorporated herein by reference, particularly for the antibody 2C4.
In an preferred embodiment of the invention, in the use of the HER dimerization inhibitor for the manufacture of a pharmaceutical composition for treating a human cancer patient, the treatment includes assessing the marker gene or the combination of marker genes at least one time or repeatedly during treatment. Preferably, the level of expression of the marker gene or the level of expression of the combination of marker genes is assessed. Preferably, the HER inhibitor is an antibody, preferably the antibody 2C4. Preferably, the patient is a breast cancer, ovarian cancer, lung cancer or prostate cancer patient.
In all embodiments of the invention, combinations of the marker genes, marker polynucleotides or marker proteins are used as disclosed above. In all embodiments of the invention, preferred values for the difference of the level of expression determined in the respective steps are also as disclosed above.
The following examples, sequence listing and figures are provided to aid the understanding of the present invention, the true scope of which is set forth in the appended claims. It is understood that modifications can be made in the procedures set forth without departing from the spirit of the invention.

EXAMPLES

Statistical Methods

The statistical tasks comprise the following steps:
1. Pre-selection of candidate biomarkers
2. Pre-selection of relevant clinical prognostic covariates
3. Selection of biomarker prediction functions at an univariate level
4. Selection of biomarker prediction functions including clinical covariates at an univariate level
5. Selection of biomarker prediction functions at a multivariate level
6. Selection of biomarker prediction functions including clinical covariates at a multivariate level
The following text details the different steps:
Ad1: Pre-selection of candidate biomarkers: The statistical pre-selection of candidate biomarkers is oriented towards the strength of association with measures of clinical benefit. For this purpose the different clinical endpoints may be transformed in derived surrogate scores, as e.g. an ordinal assignment of the degree of clinical benefit or morbidity scores regarding TTP or TTD which avoid censored observations. These surrogate transformed measures can be easily used for simple correlation analysis, e.g. by the non-parametric Spearman rank correlation approach. An alternative here is to use the biomarker measurements as metric covariates in Time-to-event regression models, as e.g. Cox proportional hazard regression. Depending on the statistical distribution of the biomarker values this step may require some pre-processing, as e.g. variance stabilizing transformations and the use of suitable scales or, alternatively, a standardization step like e.g. using percentiles instead of raw measurements. A further approach is inspection of bivariate scatter plots, e.g. by displaying the scatter of (x-axis=biomarker value, y-axis=measure of clinical benefit) on a single patient basis. Here also some non-parametric regression line as e.g. achieved by smoothing splines can be useful to visualize the association of biomarker and clinical benefit.
The goal of these different approaches is the pre-selection of biomarker candidates, which show some association with clinical benefit in at least one of the benefit measures employed, while results for other measures are not contradictory. When there are available control groups, then differences in association of biomarkers with clinical benefit in the different arms could be a sign of differential prediction which makes the biomarker eligible for further consideration.
Ad2: Pre-selection of relevant clinical prognostic covariates: The term “clinical covariate” here is used to describe all other information about the patient, which are in general available at baseline. These clinical covariates comprise demographic information like sex, age etc., other anamnestic information, concomitant diseases, concomitant therapies, result of physical examinations, common laboratory parameters obtained, known properties of the target tumor, information quantifying the extent of malignant disease, clinical performance scores like ECOG or Karnofsky index, clinical disease staging, timing and result of pretreatments and disease history as well as all similar information, which may be associated with the clinical prognosis. The statistical pre-selection of clinical covariates parallels the approaches for pre-selecting biomarkers and is as well oriented towards the strength of association with measures of clinical benefit. So in principle the same methods apply as considered under 1. In addition to statistical criteria, also criteria from clinical experience and theoretical knowledge may apply to pre-select relevant clinical covariates.
The prognosis by clinical covariates could interact with the prognosis of the biomarkers. They will be considered for refined prediction rules if necessary.
Ad3: Selection of biomarker prediction functions at an univariate level: The term “prediction function” will be used in a general sense to mean a numerical function of a biomarker measurement which results in a number which is scaled to imply the target prediction.
A simple example is the choice of the Heaviside function for a specific cutoff c and a biomarker measurement x, where the binary prediction A or B is to be made, then
If H(x−c)=0 then predict A.
If H(x−c)=1 then predict B.
This is probably the most common way of using univariate biomarker measurements in prediction rules. The definition of a prediction function usually recurs to an existing training data set which can be used to explore the prediction possibilities. In order to achieve a suitable cutoff c from the training set different routes can be taken. First the scatterplot with smoothing spline mentioned under 1 can be used to define the cutoff. Alternatively some percentile of the distribution could be chosen, e.g. the median or a quartile. Cutoffs can also be systematically extracted by investigating all possible cutoffs according to their prediction potential with regard to the measures of clinical benefit. Then these results can be plotted to allow for an either manual selection or to employ some search algorithm for optimality. This was realized based on the endpoints TTP and TTD using a Cox model, where at each test cutoff the biomarker was used as a binary covariate. Prediction criteria were the resulting Hazard ratios. Then the results for TTP and TTD can be considered together in order to chose a cutoff which shows prediction in line with both endpoints
Another uncommon approach for choosing a prediction function can be based on a fixed parameter Cox regression model obtained from the training set with biomarker values (possibly transformed) as covariate. Then the prediction could simply depend on whether the computed Hazard ratio is smaller or greater than 1.
A further possibility is to base the decision on some likelihood ratio (or monotonic transform of it), where the target probability densities were pre-determined in the training set for separation of the prediction states. Then the biomarker would be plugged into some function of the density ratios.
Ad4: Selection of biomarker prediction functions including clinical covariates at an univariate level: Univariate here refers to using only one biomarker—with regard to clinical covariates this can be a multivariate model. This approach parallels the search without clinical covariates, only that the methods should allow for incorporating the relevant covariate information. The scatterplot method of choosing a cutoff allows only a limited use of covariates, e.g. a binary covariate could be color coded within the plot. If the analysis relies on some regression approach then the use of covariates (also many of them at a time) is usually facilitated. The cutoff search based on the Cox model described under 3, allows for an easy incorporation of covariates and thereby leads to a covariate adjusted univariate cutoff search. The adjustment by covariates may be done as covariates in the model or via the inclusion in a stratified analysis.
Also the other choices of prediction functions allow for the incorporation of covariates.
This is straightforward for the Cox model choice as prediction function. There is the option to estimate the influence of covariates on an interaction level, which means that e.g. for different age groups different Hazard ratios apply.
For the likelihood ratio type of prediction functions, the prediction densities must be estimated including covariates. Here the methodology of multivariate pattern recognition can be used or the biomarker values can be adjusted by multiple regression on the covariates (prior to density estimation).
The CART technology (Classification And Regression Trees; Breiman L., Friedman J. H., Olshen R. A., Stone C. J., Chapman & Hall (Wadsworth, Inc.), New York, 1984) can be used for a biomarker (raw measurement level) plus clinical covariates employing a clinical benefit measure as response. This way cutoffs are searched and a decision tree type of functions will be found involving the covariates for prediction. The cutoffs and algorithms chosen by CART are frequently close to optimal and may be combined and unified by considering different clinical benefit measures.
Ad5: Selection of biomarker prediction functions at a multivariate level: When there are several biomarker candidates which maintain their prediction potential within the different univariate prediction function choices, then a further improvement may be achieved by combinations of biomarkers, i.e. considering multivariate prediction functions.
Based on the simple Heaviside function model combinations of biomarkers may be evaluated, e.g. by considering bivariate scatterplots of biomarker values where optimal cutoffs are indicated. Then a combination of biomarkers can be achieved by combining different Heaviside function by the logical AND and OR operators in order to achieve an improved prediction.
The CART technology (Classification And Regression Trees) can be used for multiple biomarkers (raw measurement level) and a clinical benefit measure as response, in order to achieve cutoffs for biomarkers and decision tree type of functions for prediction. The cutoffs and algorithms chosen by CART are frequently close to optimal and may be combined and unified by considering different clinical benefit measures.
The Cox-regression can be employed on different levels. A first way is to incorporate the multiple biomarkers in a binary way (i.e. based on Heaviside functions with some cutoffs). The other option is to employ biomarkers in a metric way (after suitable transformations), or a mixture of the binary and metric approach. The evolving multivariate prediction function is of the Cox type as described under 3.
The multivariate likelihood ratio approach is difficult to realize but presents as well as an option for multivariate prediction functions.
Ad6: Selection of biomarker prediction functions including clinical covariates at a multivariate level: When there are relevant clinical covariates then a further improvement may be achieved by combining multiple biomarkers with multiple clinical covariates. The different prediction function choices will be evaluated with respect to the possibilities to include clinical covariates.
Based on the simple logical combinations of Heaviside functions for the biomarkers, further covariates may be included to the prediction function based on logistic regression model obtained in the training set.
The CART technology and the evolving decision trees can be easily used with additional covariates, which would include these in the prediction algorithm.
All prediction functions based on the Cox-regression can use further clinical covariates. There is the option to estimate the influence of covariates on an interaction level, which means that e.g. for different age groups different Hazard ratios apply.
The multivariate likelihood ratio approach is not directly extendible to the use of additional covariates.

Example 1

Baseline Blood Sera from HER2 Low Expressing Metastatic Breast Cancer Patients Treated with Pertuzumab were Assessed for Levels of HER Ligands and Shedded HER2 (HER2 ECD), as Described Below

Kits used for assessment of the serum biomarkers:


Marker	Assay	Distribution

HER2-ECD	Bayer HER-2/neu ELISA,	DakoCytomation N.V./S.A.,
	Cat. #: EL501	Interleuvenlaan 12B,
		B-3001 Heverlee
Amphiregulin	DuoSet ELISA Development	R&D Systems Ltd., 19 Barton
	System Human Amphiregulin,	Lane, Abingdon OX14 3NB, UK
	Cat. #: DY262
EGF	Quantikine human EGF	R&D Systems Ltd., 19 Barton
	ELISA kit, Cat. #: DEG00	Lane, Abingdon OX14 3NB, UK
TGF-alpha	Quantikine ® Human TGF-alpha	R&D Systems Ltd., 19 Barton
	Immunoassay, Cat. #: DTGA00	Lane, Abingdon OX14 3NB, UK

Protocols:

HER2-ECD:

HER2-ECD ELISA was performed according to the recommendations of the manufacturer.
Amphiregulin:

- Prepare all reagents (provided with the kit), standard dilutions (provided with the kit) and samples

Provide EvenCoat Goat Anti-mouse IgG microplate strips (R&D, Cat. # CP002; not provided with the kit) in the frame. The frame is now termed ELISA plate.
Determine of the required number of wells (number of standard dilutions+number of samples).
Determine the plate layout.
Add 100 μl diluted capture antibody (provided with the kit; 1:180 in PBS) to each well.
Incubate at r.t. for 1 hour.
Aspirate each well and wash, repeating the process three times for a total of four washes. Wash by filling each well with 400 μl Wash buffer (not provided with the kit; 0.05% Tween-20 in PBS was used), using a manifold dispenser, and subsequent aspiration. After the last wash, remove any remaining Wash buffer by aspirating. Invert the plate and blot it against clean paper towels.
Add 100 μl standard dilution or diluted sample (see below) per well. Change tip after every pipetting step.
Cover plate with the adhesive strip (provided with the kit).
Incubate for 2 hours at r.t. on a rocking platform.
Repeat the aspiration/wash as described previously.
Aspirated samples and wash solutions are treated with laboratory disinfectant.
Add 100 μl Detection Antibody (provided with the kit) diluted 1:180 in Reagent diluent (not provided with the kit; 1% BSA (Roth; Albumin Fraction V, Cat. # T844.2) in PBS was used) per well
Incubate for 2 hours at r.t.
Repeat the aspiration/wash as described previously.
Add 100 μl working dilution of the Streptavidin-HRP to each well (provided with the kit; 1:200 dilution in Reagent diluent). Cover with a new adhesive strip.
Incubate for 20 min at r.t.
Repeat the aspiration/wash as described previously.
Add 100 μl Substrate Solution (R&D, Cat. # DY999; not provided with the kit) to each well.
Incubate for 20 min at r.t. Protect from light. Add 50 μl Stop Solution (1.5 M H2504 (Schwefelsäure reinst, Merck, Cat. # 713); not provided with the kit) to each well. Mix carefully.
Determine the optical density of each well immediately, using a microplate reader set to 450 nm.

Amphiregulin Standard Curve:

A 40 ng/ml amphiregulin stock solution was prepared in 1% BSA in PBS, aliquotted and stored at −80° C. Amphiregulin solutions in 20% BSA in PBS were not stable beyond 2 weeks and were therefore not used. From the aliquotted amphiregulin stock solution, the amphiregulin standard curve was prepared freshly in 20% BSA in PBS prior to each experiment. The highest concentration was 1000 pg/ml (1:40 dilution of the amphiregulin stock solution). The standards provided with the ELISA kit produced a linear standard curve. Excel-based analysis of the curves allowed the determination of curve equations for every ELISA.

Amphiregulin Samples:

When samples were diluted 1:1 in Reagent Diluent, all samples were within the linear range of the ELISA. Each sample was measured in duplicates. Dependent on the quality of the data, and on sufficient amounts of serum, determinations were repeated in subsequent experiments if necessary.
EGF:
Prepare all reagents (provided with the kit), standard dilutions (provided with the kit) and samples
Remove excess antibody-coated microtiter plate strips (provided with the kit) from the frame. The frame is now termed ELISA plate.
Determine of the required number of wells: (Number of standard dilutions+number of samples)×2
Determine the plate layout.
Add 50 μl Assay Diluent RD1 (provided with the kit) to each well
Add 200 μl standard dilution or diluted sample (e.g. 1:20 in Calibrator Diluent RD6H) per well. Change tip after every pipetting step.
Cover plate with the adhesive strip (provided with the kit).
Incubate for 2 hours at r.t. on a rocking platform.
Aspirate each well and wash, repeating the process three times for a total of four washes. Wash by filling each well with 400 μl Wash Buffer (provided with the kit), using a manifold dispenser, and subsequent aspiration. After the last wash, remove any remaining Wash buffer by aspirating. Invert the plate and blot it against clean paper towels.
Aspirated samples and wash solutions are treated with laboratory disinfectant. Add 200 μl of Conjugate (provided with the kit) to each well. Cover with a new adhesive strip.
Incubate for 2 hours at r.t.
Repeat the aspiration/wash as described previously.
Add 200 μl Substrate Solution (provided with the kit) to each well.
Incubate for 20 min at r.t. Protect from light.
Add 50 μl Stop Solution (provided with the kit) to each well. Mix carefully.
Determine the optical density of each well within 30 minutes, using a microplate reader set to 450 nm.

EGF Standard Curve:

The standards provided with the ELISA kit produced a linear standard curve. Also very small concentrations showed detectable results.
EGF samples:
A total of four assays with the samples was performed. Each sample was measured 2-5 times, the number of determinations being dependent on the quality of the results (mean+/−SD) and the availability of sufficient amounts of serum. When samples were diluted 1:20 in Calibrator Diluent RD6H, all samples were within the linear range of the ELISA.

TGF-Alpha:

Prepare all reagents (provided with the kit), standard dilutions (provided with the kit) and samples
Remove excess antibody-coated microtiter plate strips (provided with the kit) from the frame. The frame is now termed ELISA plate.
Determine of the required number of wells: (Number of standard dilutions+number of samples)×2
Determine the plate layout.
Add 100 μl Assay Diluent RD1W (provided with the kit) to each well
Add 50 μl standard dilution or sample per well. Change tip after every pipetting step.
Cover plate with the adhesive strip (provided with the kit).
Incubate for 2 hours at r.t. on a rocking platform.
Aspirate each well and wash, repeating the process three times for a total of four washes. Wash by filling each well with 400 μl Wash Buffer (provided with the kit), using a manifold dispenser, and subsequent aspiration. After the last wash, remove any remaining Wash buffer by aspirating. Invert the plate and blot it against clean paper towels.
Aspirated samples and wash solutions are treated with laboratory disinfectant.
Add 200 μl of TGF-alpha Cojugate (provided with the kit) to each well. Cover with a new adhesive strip.
Incubate for 2 hours at r.t.
Repeat the aspiration/wash as described previously.
Add 200 μl Substrate Solution (provided with the kit) to each well.
Incubate for 30 min at r.t. Protect from light.
Add 50 μl Stop Solution (provided with the kit) to each well. Mix carefully.
Determine the optical density of each well within 30 minutes, using a microplate reader set to 450 nm.

TGF-Alpha Standard Curve:

The standards provided with the ELISA kit produced a linear standard curve. Also very small concentrations showed detectable results.

TGF-Alpha Samples:

A total of four assays with the samples was performed. Samples were measured in 2-4 independent assays.
The serum data was analyzed to identify factors the baseline serum levels of which would be associated with response to the Pertuzumab treatment. For all factors a skewed pattern of the distribution (mean, standard deviation, median, minimum, maximum) was observed. A monotonic transform was used to reduce the skewness based on the logarithm: Log(x+1). In a univariate analysis, it was explored whether suitable cut-points for the factors could be defined which would relate to the probability of response (in this example defined as clinical benefit). Here, patients with clinical benefit were defined as those who achieved a partial response (PR) or maintained stable disease for at least 6 months. Scatterplots of the factors versus the response categories were investigated. FIG. 1 and FIG. 2 show a plotting of the clinical response categories versus the logarithmic transformation of the serum levels of TGF-alpha and amphiregulin, respectively, to exemplify the approach.
Based on the scatterplots, cut-points were selected for the factors to define groups of patients, who have experienced greater clinical benefit. FIG. 3 (TGF-alpha), FIG. 4 (Amphiregulin), FIG. 5 (EGF), and FIG. 6 (HER2-ECD) show the clinical benefit in relation to the different factor groupings based on the exploratory cut-points calculated to the original factor units. The cut-points separate out some of the patients without clinical benefit, and hence, elevate the response rate for the group with greater clinical benefit.

Example 2

In this example the exploratory cut-points from Example 1 were used to assess the univariate effect of the factor groupings on different measures of the clinical benefit of the Pertuzumab treatment, using time to progression/or death (TTP) and time to death (TTD) as alternative clinical endpoints. Significant effects were observed for TGF-alpha, Amphiregulin, EGF and HER2-ECD in Kaplan-Meier estimates and log-rank tests for TTP and/or TTD, as shown in an overview in FIG. 7.
The Kaplan-Meier plots displaying the hazard ratio are given for TTP and TTD (highest number of events observed) in FIG. 8 and FIG. 9 (TGF-alpha), 10 and 11 (Amphiregulin), 12 and 13 (EGF), and 14 and 15 (HER2-ECD), showing the pronounced effect of a grouping based on these factors on the clinical outcome of the patients treated with Pertuzumab.

Example 3

In this example multivariate approaches were used to identify combinations of factors that would further improve the identification of patients with greater benefit from the Pertuzumab treatment. Results, as derived from a CART approach (Classification And Regression Trees), are reflected. The CART classification approach made it necessary to specify as the benefit group all values in clinical benefit above of 0. As variables serum levels of HER2-ECD, TGF-alpha, Amphiregulin, and EGF were employed. A combination of serum HER2-ECD and serum TGF-alpha levels were selected to give best results. From the CART results optimized cut-points for a combination of serum HER2-ECD and serum TGF-alpha levels were derived, resulting in a rule for exploratory categorization of clinical benefit in the study population—a combination of low serum HER2-ECD values and low serum TGF-alpha values capturing 2/2 PR and 2/3 SD>6 months in the study population and excluding a reasonable number of fast progressing patients. FIG. 16 shows the clinical benefit in relation to the TGF-alpha/HER2-ECD combination groupings based on the exploratory combination cut-point. FIG. 17 summarizes the effect of a combination of TGF-alpha and HER2-ECD on TTP and TTD. The Kaplan-Meier estimates and the hazard ratios given in FIG. 18 (TTP) and FIG. 19 (TTD) demonstrate the significant effect of the grouping based on a combination of these factors for on the clinical outcome of the patients treated with Pertuzumab.
Unless stated to the contrary, all compounds in the examples were prepared and characterized as described. All ranges recited herein encompass all combinations and subcombinations included within that range limit. All patents and publications cited herein are hereby incorporated by reference in their entirety for any purpose.

Claims

1. A method of predicting the response of a patient to treatment with a HER dimerization inhibitor comprising the steps of:

(a) determining the expression level one or more biomarker proteins with one such biomarker protein being transforming growth factor alpha;

(b) determining whether the expression level assessed in step (a) is above or below a certain quantity that is associated with an increased or decreased clinical benefit to a patient; and

(c) predicting the response to the treatment with the HER inhibitor in the patient by evaluating the results of step (b).

2. A method according to claim 1 wherein, in step (a), in addition to the expression level of transforming growth factor alpha, the level of expression of one or more biomarker proteins, selected from the group consisting of epidermal growth factor, amphiregulin, and HER2, is also determined.

3. A method according to claim 2 wherein, in step (a), the level of expression of transforming growth factor alpha is determined in combination with the level of expression of: epidermal growth factor; HER2; or both amphiregulin and HER2.

4. A method according to claim 1 wherein the HER dimerization inhibitor inhibits heterodimerization of HER2 with EGFR or HER3.

5. A method according to claim 1 wherein the HER dimerization inhibitor is an antibody.

6. A method according to claim 1 wherein the HER dimerization inhibitor is the antibody 2C4.

7. A method according to claim 1 wherein the patient is a cancer patient.

8. A method according to claim 1 wherein the patient is a breast cancer, ovarian cancer, lung cancer or prostate cancer patient.

9. A method according to claim 1 wherein the level of expression of said biomarker protein or proteins is assessed by detecting the level of expression of fragments thereof.

10. A method according to claim 1 wherein the level of expression said biomarker or biomarkers is determined using a reagent which specifically binds with the biomarker protein or the fragment thereof or the combination of biomarker proteins or the fragments thereof.

11. A method according to claim 10 wherein said reagent is selected from the group consisting of an antibody, a fragment of an antibody, and an antibody derivative.

12. A method according to claim 1 wherein a biomarker protein is the extracellular domain of HER2.

13. A method according to claim 12 wherein said extracellular domain extracellular domain of HER2 has a molecular mass of approximately 105,000 Daltons.

14. A method according to claim 1 wherein the amino acid sequence of the transforming growth factor alpha biomarker protein is SEQ ID NO: 3

15. A method according to claim 2 wherein the amino acid sequence of the amphiregulin biomarker protein is SEQ ID NO: 1, the amino acid sequence of the epidermal growth factor biomarker protein is SEQ ID NO: 2, the amino acid sequence of the transforming growth factor alpha biomarker protein is SEQ ID NO: 3, and the amino acid sequence of the HER2 biomarker protein is SEQ ID NO: 4.

16. A method according to claim 1 wherein the amount of transforming growth factor alpha is determined in a biological sample of blood serum and the quantity that is associated with an increased clinical benefit to a patient is within the range between 2.0 to 10.0 pg/ml of transforming growth factor alpha.

17. A method according to claim 2 wherein the amounts of transforming growth factor alpha, amphiregulin, epidermal growth factor, and HER2 marker proteins are determined in a biological sample of blood serum and the quantities that are associated with an increased clinical benefit to a patient are within the range of between 2.0 to 10.0 pg/ml of transforming growth factor alpha, between 6 to 15 pg/ml of amphiregulin, between 100 to 250 pg/ml of epidermal growth factor, and between 12 to 22 ng/ml of HER2 marker proteins.

18. A method according to claim 2 wherein the amounts of transforming growth factor alpha, amphiregulin, epidermal growth factor, and HER2 marker proteins are determined in a biological sample of blood serum and the quantities that are associated with an increased clinical benefit to a patient are about 3.5 pg/ml of transforming growth factor alpha, about 12 pg/ml of amphiregulin, about 150 pg/ml of epidermal growth factor, and about 18 ng/ml of HER2 marker proteins.