Scoring method for anti-FRα antibody drug conjugate therapy

Abstract

A method for predicting the efficacy of antibody drug conjugate (ADC) therapy involves determining the average optical density (OD) of the membrane and, optionally, the cytoplasmic staining of each cancer cell in a whole slide image. The diagnostic antibody to which the dye is linked targets the folate receptor alpha protein in the same way as the ADC antibody. Statistical calculations are performed on the OD of the staining, such as calculating the median, absolute deviation, the difference in OD between membrane and cytoplasmic staining, the percentage of cells with the lowest OD, or the sum of minimally stained cells and poorly stained cells near the stained cells. For example, based on whether the median OD is below, between, or above two OD thresholds, a zero, high, or low ADC dose is recommended, respectively. The upper and lower OD thresholds are correlated with the response of training patients who were treated with the ADC and whose stained cancer tissues were analyzed.

Description

[Technical field]

[0001] The present invention relates to methods for determining a score that indicates how a cancer patient will respond to therapy using an anti-FRα antibody drug conjugate. [Background technology]

[0002] Evaluating the probability of a cancer patient's response to a given treatment is an essential step in determining the treatment regimen for the cancer patient. Such evaluation is often based on histological analysis of tissue samples from cancer patients and involves identifying and classifying the cancer using a standard grading scheme. Immunohistochemistry (IHC) staining can be used to distinguish marker-positive cells that express a particular protein from marker-negative cells that do not express that protein. IHC staining typically involves multiple dyes, including one or more dyes attached to a protein-specific antibody and another dye that is a counterstain. A common counterstain is hematoxylin, which labels DNA and therefore stains the nucleus.

[0003] Protein-specific stains or biomarkers can be used to identify areas of a cancer patient's tissue (e.g., tumor tissue) that are likely to respond to a given therapy. For example, biomarkers that stain epithelial cells can help identify suspicious tumor areas. Other protein-specific biomarkers are then used to characterize cells within the cancerous tissue. Cells stained by a particular biomarker can be identified and quantified, and then a score indicating the number of positively and negatively stained cells can be visually estimated by a pathologist. This score can then be compared to scores of other cancer patients calculated in the same way. If the response of these other patients to a given cancer treatment is known, the pathologist can predict how likely the cancer patient is to respond to a given treatment based on a comparison of the score calculated for the cancer patient with the scores of the other patients. However, visual assessment by a pathologist tends to be variable and subjective.

[0004] One promising cancer treatment involves antibody-drug conjugates (ADCs), which have a cytotoxic drug conjugated to an antibody whose antigen is expressed on the surface of cancer cells. The ADCs bind to the antigen and undergo cellular internalization, selectively delivering the drug to the cancer cells, where it accumulates and kills them.

[0005] Folate receptors (FRs) are membrane-bound proteins present on the cell surface and can therefore be utilized to develop new ADCs. The FR family includes FRα, FRβ, FRγ, and FRδ. FRs bind folate molecules and transport them into cells, thereby delivering them to the folate cycle to support nucleotide metabolism. In particular, folate is important for DNA synthesis, methylation, and repair (Cheung, et al., Oncotarget. 2016; 7(32):52553-52574).

[0006] FRα is a glycosylphosphatidylinositol (GPI)-anchored membrane protein with high affinity for the active form of folate, 5-methyltetrahydrofolate (5-MTF). Previous studies have shown that FRα plays an important role in embryonic development (Kelemen, Int J Cancer. 2006; 119(2): 243-250). However, folate transport in adults is mainly driven by the ubiquitous expression of reduced folate carriers and proton-conjugated folate transporters (Zhao, et al., Annu Rev Nutr. 2011; 31: 177-201). The distribution of FRα expression in adults is usually restricted to the apical membrane side of polarized epithelia, such as the choroid plexus, kidney, lung, and placenta.

[0007] Overexpression of FRα, also known as folate-binding protein (FBP) and associated with the gene FOLR1, is frequently observed in ovarian, lung (e.g., non-small cell lung cancer (NSCLC)), and breast cancer (Shi, et al., Drug Des Devel Ther. 2015; 9: 4989-4996). Notably, previous studies have found elevated levels of soluble FRα in the blood of ovarian cancer patients, supporting the potential application of FRα as a biomarker for early ovarian cancer (Basal, et al., PLoS One. 2009; 4(7): e6292). Preclinical ovarian models have also revealed that overexpression of FRα is associated with tumor progression and that binding of folate to FRα may mediate activation of the proto-oncogene STAT3 (Hansen, et al., Cell Signal. 2015; 27(7): 1356-1368). Summary of the Invention

[0008] There is a need for computer-based methods to generate reproducible, objective scores that predict a cancer patient's response to treatment with therapeutic FRα antibody drug conjugates.

[0009] A method for predicting how cancer patients will respond to therapy involving antibody drug conjugates (ADCs) involves determining a predictive efficacy score based on membrane optical density and optionally cytoplasmic staining by a dye linked to a diagnostic antibody. The ADC comprises an ADC payload and an ADC antibody that targets a protein on each cancer cell. Both the diagnostic antibody and the ADC antibody target the folate receptor alpha (FRα) protein on the cancer cell.

[0010] A tissue sample is immunohistochemically stained using a dye linked to a diagnostic antibody that binds to FRα protein on cancer cells in the tissue sample. A digital image of the tissue sample is obtained. Image analysis is performed on the digital image to detect cancer cells using a convolutional neural network. For each cancer cell, the optical density of the stain is determined based on the staining intensity of the dye in the membrane and / or cytoplasm of the cancer cell, and / or in the membrane and cytoplasm of other cancer cells that are closer to the cancer cell than a predetermined distance. Based on various statistical operations performed on the optical density of the stain, a predictive efficacy score is generated that predicts the response of the cancer patient to ADC therapy. Patients with a score higher than a predetermined threshold are recommended therapy with ADC.

[0011] In one embodiment, a method for predicting a cancer patient's response to an anti-FRα ADC involves detecting cancer cells and determining the mean optical density (optionally the median of the mean optical density) of membrane staining of all cancer cells in a digital image of a tissue sample from the cancer patient. The ADC includes an ADC payload and an ADC antibody that targets the FRα protein on the cancer cells. A dye linked to a diagnostic antibody is used to immunohistochemically stain the tissue sample. The diagnostic antibody binds to the FRα protein on the cancer cells in the tissue sample. A digital image of the tissue sample is obtained and the cancer cells are detected in the digital image. For each cancer cell, the mean staining intensity of the dye in the membrane is measured. In one embodiment, the percentage of cancer cells having a staining intensity greater than a predetermined threshold is measured. In another embodiment, the staining intensity of the dye in the membrane of other cancer cells closer to the cancer cell than a predetermined distance is also measured. The staining intensity of each membrane is calculated based on the mean optical density of the brown diaminobenzidine (DAB) signal at the pixels of the membrane. The cancer patient's response to the ADC is predicted based on whether the predicted efficacy score exceeds a predetermined threshold. Therapy involving the ADC may be administered to a cancer patient if the predicted efficacy score exceeds a predetermined threshold.

[0012] In an embodiment relating to a median optical density above a threshold, a predicted efficacy score is generated to predict a cancer patient's response to an ADC comprising an ADC payload and an ADC antibody targeting the folate receptor alpha (FRα) protein on cancer cells. A cancer patient's tissue sample is immunohistochemically stained using a dye linked to a diagnostic antibody that binds to the FRα protein on cancer cells in the tissue sample. A digital image of the tissue sample is acquired. For each cancer cell, the average optical density of the dye staining at the membrane of the cancer cell is determined. Then, the median optical density of the staining of all cancer cells in the digital image is determined. A predicted efficacy score is generated for the tissue sample based on the median optical density. The predicted efficacy score is positive if the median optical density is equal to or greater than the optical density threshold, and is negative if the median optical density is less than the optical density threshold. The optical density threshold correlates with the response of a cohort of training patients treated with the ADC. A therapy involving the ADC is recommended for the cancer patient if the predicted efficacy score is positive. A therapy involving the ADC may be administered to the cancer patient if the predicted efficacy score is positive.

[0013] In an embodiment relating to the percentage of cancer cells with a mean optical density of staining above a minimum optical density threshold, a predictive efficacy score is generated that predicts a cancer patient's response to an ADC comprising an ADC payload and an ADC antibody that targets an FRα protein on cancer cells. A tissue sample is immunohistochemically stained using a dye linked to a diagnostic antibody that binds to a protein on cancer cells in the tissue sample. A digital image of the tissue sample is obtained. Cancer cells in the digital image are detected using image analysis. For each cancer cell, the optical density of the dye staining at the cell membrane is determined. Each cancer cell is identified as either (i) optical density positive if the optical density of the cancer cells is equal to or greater than the optical density threshold, or (ii) optical density negative if the mean optical density of the cancer cells is less than the optical density threshold. A predictive efficacy score is generated for the tissue sample based on the percentage of cancer cells in the digital image that are optical density positive. The predicted efficacy score is positive if the percentage of cancer cells that are optical density positive is equal to or greater than the percentage threshold, and is negative if the percentage of cancer cells that are optical density positive is less than the percentage threshold. The optical density threshold and the percentage threshold are correlated with the response of a cohort of training patients treated with ADC. Therapy with ADC is recommended for the cancer patient if the predicted efficacy score is positive. Therapy with ADC can be administered to the cancer patient if the predicted efficacy score is positive.

[0014] In an embodiment relating to the spatial proximity of differently stained cells, a cancer patient is identified for treatment with an ADC comprising an ADC payload and an ADC antibody targeting the FRα protein on cancer cells. A tissue sample from the cancer patient is immunohistochemically stained using a dye linked to a diagnostic antibody that binds to the FRα protein on cancer cells in the tissue sample. A digital image of the tissue sample is obtained. Cancer cells are detected in the digital image. For each cancer cell, the average optical density of the dye staining at the cell membrane is determined. Each cancer cell is identified as either (i) optical density positive if the average optical density of the cancer cells is equal to or greater than an optical density threshold, or (ii) optical density negative if the average optical density of the cancer cells is less than an optical density threshold. A proximity score is generated for the tissue sample, which is equal to the percentage of cancer cells in the digital image that are either optical density positive or optical density negative, but within a predetermined distance from the optical density positive cancer cells. A cancer patient is identified as one who would likely benefit from administration of the ADC if the proximity score exceeds a predetermined percentage threshold. Therapy involving ADC may be administered to a cancer patient if the proximity score exceeds a certain percentage threshold.

[0015] In an embodiment relating to the median absolute deviation of the optical density of the staining of cancer cells in the whole slide image, a predictive efficacy score is generated that predicts the response of a cancer patient to an ADC that includes an ADC payload and an ADC antibody that targets the FRα protein on the cancer cells. A tissue sample is immunohistochemically stained using a dye linked to a diagnostic antibody that binds to the FRα protein on the cancer cells in the tissue sample. A digital image of the tissue sample is obtained. The cancer cells are detected in the digital image. For each cancer cell, the average optical density of the staining by the dye in the membrane of the cancer cell is determined. The median optical density of all cancer cells in the digital image is then determined. From the median optical density of all cancer cells in the digital image, the median absolute deviation of the optical density of the cancer cells is determined. A predictive efficacy score for the tissue sample is generated based on the median absolute deviation. The predictive efficacy score is positive if the median absolute deviation is equal to or greater than a deviation threshold, and is negative if the median absolute deviation is less than a deviation threshold. The deviation threshold correlates with the response of a cohort of training patients treated with the ADC. Therapy involving ADC is recommended for the cancer patient if the predicted efficacy score is positive. Therapy involving ADC may be administered to the cancer patient if the predicted efficacy score is positive.

[0016] In an embodiment relating to recommending an effective dose of an ADC based on membrane staining, a recommended dose is determined for treating a cancer patient with an ADC comprising an ADC payload and an ADC antibody targeting an FRα protein on cancer cells. A tissue sample is immunohistochemically stained using a dye linked to a diagnostic antibody that binds to the FRα protein on cancer cells in the tissue sample. A digital image of the tissue sample is obtained. Cancer cells in the digital image are detected. For each cancer cell, the average optical density of the dye staining in the membrane of the cancer cell is determined. The median optical density of all cancer cells in the digital image is then determined. A recommended dose is determined based on whether the median optical density is below a lower optical density threshold, between a lower optical density threshold and an upper optical density threshold, or above an upper optical density threshold. The recommended dose is zero if the median optical density is below a lower optical density threshold, a higher dose if the median optical density is between a lower optical density threshold and an upper optical density threshold, and a lower dose if the median optical density is above an upper optical density threshold. The lower and upper optical density thresholds are correlated with the response of a cohort of training patients treated with ADC. Using this method, treatment recommendations and dosage recommendations involving ADC are made to cancer patients. A recommended dosage of therapy involving ADC can be administered to cancer patients based on the median optical density of all cancer cells in the digital image.

[0017] In an embodiment relating to the difference between the optical density of the membrane staining and the optical density of the cytoplasmic staining, a predictive efficacy score is generated that predicts the response of a cancer patient to an ADC that includes an ADC payload and an ADC antibody that targets the FRα protein on cancer cells. A tissue sample is immunohistochemically stained using a dye linked to a diagnostic antibody that binds to a protein on the membrane and in the cytoplasm of cancer cells in the tissue sample. A digital image, such as a whole slide image, of the tissue sample is obtained. Cancer cells in the digital image are detected using image analysis, optionally involving a convolutional neural network. For each cancer cell, the average optical density of the dye staining in the membrane of the cancer cell is determined. For each cancer cell, the average optical density of the dye staining in the cytoplasm of the cancer cell is determined. The difference between the average optical density of the membrane staining and the average optical density of the cytoplasmic staining is determined for each cancer cell in the digital image. The 85% quantile of the difference between the average optical density of the membrane staining and the average optical density of the cytoplasmic staining is identified among all cancer cells in the digital image. If the 85% quantile of the difference exceeds a predetermined difference threshold, a therapy with ADC is recommended for the cancer patient. If the 85% quantile of the difference exceeds a predetermined difference threshold, a therapy with ADC can be administered to the cancer patient.

[0018] Other embodiments and advantages are described in the following detailed description. This summary is not intended to define the invention, which is defined by the claims.

[0019] The accompanying drawings, in which like numerals indicate like elements, illustrate embodiments of the present invention. [Brief description of the drawings]

[0020] [Figure 1] 1 shows the amino acid sequences of the various CDRs of various constructs of ADC anti-FRα antibodies (SEQ ID NO: 1 to SEQ ID NO: 36). [Diagram 2] 1 shows the amino acid sequences of the heavy and light chain variable regions of various constructs of anti-FRα antibodies (SEQ ID NO: 37 to SEQ ID NO: 48). [Diagram 3]The amino acid sequences of the heavy and light chains of the three constructs of the anti-FRα antibody (SEQ ID NO: 49 to SEQ ID NO: 54) are shown, respectively. [Figure 4] The amino acid sequences of the heavy and light chains of three additional constructs of anti-FRα antibody (SEQ ID NO: 55 to SEQ ID NO: 60) are shown, respectively. [Diagram 5] 1 shows the amino acid sequences of various light and heavy chain variable framework regions of anti-FRα antibodies (SEQ ID NO: 61 to SEQ ID NO: 108). [Figure 6] The amino acid sequences of the three constant domain regions of the anti-FRα antibody (SEQ ID NO: 109 to SEQ ID NO: 111) are shown. [Figure 7] 1 is a table showing three exemplary numbering systems used to identify sequences to define complementarity determining regions (CDRs). [Figure 8] The amino acid sequence of the FRα protein (SEQ ID NOs: 112 to 113) is shown. [Figure 9] 1 is a flow chart of steps in which an analysis system analyzes digital images of stained tissue to predict how a cancer patient is likely to respond to an anti-FRα ADC based on the percentage of cancer cells that are positively stained or that are not positively stained but are located near positively stained cells. [Figure 10] 10 shows digital images illustrating the image analysis process of step 12 of FIG. 9. [Figure 11] 1 illustrates the image analysis steps in which nuclear objects of cancer cells are detected. [Figure 12] FIG. 13 shows the image analysis steps used to detect membranes of nuclear objects. [Figure 13] 13 is a screenshot of the results of the image analysis step in an image analysis software environment. [Figure 14] 1 shows an example calculation of binary spatial proximity scores of 10 exemplary cells based on cell separation to reflect uptake of ADC payload into neighboring cells. [Figure 15] 1 illustrates the mechanism by which anti-FRα ADC therapy kills cancer cells. [Figure 16]Shown is a sample quantification of staining intensity from image analysis using membrane and cytoplasmic pixel grey values. [Figure 17] Exemplary quantitative staining amounts for staining at the membrane and cytoplasm of the images in FIG. 16 are listed. [Figure 18] Calculation of a continuous spatial proximity score for each of the three cells shown in FIG. 17 based on cell separation to reflect uptake of the ADC payload into adjacent cells is shown. [Figure 19] 1 shows plots and bar graphs correlating binary spatial proximity scores with actual outcomes of patient-derived xenograft (PDX) studies of patients with ovarian cancer, NSCLC, colon cancer, and endometrial cancer. [Figure 20] 13 is a plot showing binary spatial proximity scores by increasing size for 28 PDX models of ovarian cancer, NSCLC, and endometrial cancer. [Figure 21] Receiver operating characteristic (ROC) curves used to determine optimal cut-points for stratifying responsive and non-responsive PDX models. [Figure 22] 1 is a bar graph of binary spatial proximity scores ordered by magnitude for 197 tissue samples from an ovarian cancer patient database. [Figure 23] 1 is a flow chart of method steps for predicting efficacy of an anti-FRα ADC based on the percentage of cancer cells in a whole slide image that exhibits at least a minimal amount of staining. [Figure 24] 1 shows plots and bar graphs correlating predicted efficacy scores based on the percentage of minimally stained cells with actual outcomes of PDX studies in patients with ovarian, NSCLC, colon, and endometrial cancer. [Diagram 25] 1 is a plot showing predicted efficacy scores (based on % minimally stained cells) by increasing size for 28 PDX models from ovarian cancer, NSCLC, and endometrial cancer patients. [Figure 26]FIG. 2 is a ROC curve used to determine optimal cut-points for stratifying responsive and non-responsive PDX models based on the percentage of minimally stained cells. [Figure 27] 1 is a bar graph of the predictive efficacy scores (based on % minimally stained cells) ordered from highest to lowest for 197 tissue samples from a database of ovarian cancer patients. [Figure 28] 1 is a flow chart of method steps for predicting efficacy of anti-FRα ADCs based on the median absolute deviation of staining of each cell from the median membrane staining of all cancer cells in a digital image. [Figure 29] 1 shows plots and bar graphs correlating predicted efficacy scores based on median absolute deviation of staining intensity with actual outcomes of PDX studies in patients with ovarian, NSCLC, colon, and endometrial cancer. [Diagram 30] Plot showing predicted efficacy scores (median absolute deviation of staining) by increasing size of 28 PDX models from ovarian cancer, NSCLC, and endometrial cancer patients. [Diagram 31] FIG. 2 is a ROC curve used to determine optimal cut-points for stratifying responsive and non-responsive PDX models based on median absolute deviation of staining. [Diagram 32] 1 is a bar graph of ascending predictive validity scores (median absolute deviation of staining) for 197 tissue samples from a database of ovarian cancer patients. [Diagram 33] 1 is a flow chart of method steps for determining a recommended dosage of an anti-FRα ADC based on the median optical density of membrane staining in digital images. [Diagram 34] 1 shows plots demonstrating the correlation between median optical density of membrane staining and tumor response of PDX models from ovarian cancer, NSCLC, and endometrial cancer patients to anti-FRα ADC at 5 mg / kg. [Diagram 35]1 is a plot showing predicted efficacy scores (based on median optical density of membrane staining) by increasing size of 28 PDX models from ovarian cancer, NSCLC, and endometrial cancer patients for 5 mg / kg anti-FRα ADC. [Diagram 36] 2 is a ROC curve used to determine optimal cut-points for stratifying responsive and non-responsive PDX models with 5 mg / kg anti-FRα ADC based on median optical density of membrane staining. [Figure 37] 1 shows plots demonstrating the correlation between median optical density of membrane staining and tumor response of PDX models from ovarian cancer, NSCLC, and endometrial cancer patients to 2.5 mg / kg anti-FRα ADC. [Figure 38] 1 is a plot showing predicted efficacy scores (based on median optical density of membrane staining) by increasing size of 28 PDX models from ovarian cancer, NSCLC, and endometrial cancer patients for the 2.5 mg / kg anti-FRα ADC. [Figure 39] 2 is a ROC curve used to determine optimal cut-points for stratifying 2.5 mg / kg anti-FRα ADC responsive and non-responsive PDX models based on median optical density of membrane staining. [Diagram 40] Graph showing similar distribution of FRα protein in ovarian cancer PDX models and human ovarian tumors. [Diagram 41] 1 is a bar graph of predicted efficacy scores (based on median optical density of membrane staining) in ascending order for 197 ovarian cancer patients showing upper and lower optical density thresholds that stratify patients receiving higher, lower, and zero doses of anti-FRα ADC. [Diagram 42] 1 is a flow chart of method steps for predicting the efficacy of an anti-FRα ADC based on the difference between the optical density of membrane and cytoplasmic staining. [Diagram 43]1 shows plots correlating predicted efficacy scores based on the difference between membranous and cytoplasmic staining intensity with actual outcomes of PDX studies in patients with ovarian, NSCLC, colon, and endometrial cancer. [Diagram 44] 1 is a plot showing predicted efficacy scores (difference between membrane and cytoplasmic OD) by increasing size of 28 PDX models from ovarian cancer, NSCLC, and endometrial cancer patients. [Diagram 45] FIG. 2 is a ROC curve used to determine optimal cut-points for stratifying responsive and non-responsive PDX models based on the difference between membrane and cytoplasmic staining intensity. [Figure 46] 1 is a bar graph of ascending predictive validity scores (difference between membrane and cytoplasmic OD) for 197 tissue samples from a database of ovarian cancer patients. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0021] The present invention relates to a novel method for predicting the response of a cancer patient to an antibody drug conjugate (ADC) comprising an ADC antibody targeting the folate receptor alpha (FRα) protein on cancer cells, where the response is predicted based on a statistical calculation using the measured optical density of staining with a dye linked to a diagnostic antibody that also targets the FRα protein. Another aspect of the present invention relates to a method for treating a cancer patient by administering a therapy with an ADC based on a quantitative continuous score (QCS) calculated using one of the statistical calculations. Yet another aspect of the present invention relates to a method for identifying a cancer patient for treatment with an ADC based on a QCS score.

[0022] The first embodiment involves a novel method for determining a recommended dosage of an FRα antibody drug conjugate (ADC) for a cancer patient based on whether the median optical density of all cancer cells in a tissue sample stained with a diagnostic antibody that binds to a protein in the same way as the ADC antibody is below, between, or above two optical density thresholds determined based on the response of a cohort of training patients treated with the ADC.

[0023] A second embodiment relates to a method for generating a predictive efficacy score that indicates how a cancer patient will respond to therapy involving an ADC based on the median optical density of all cancer cells in a cancer patient's tissue sample stained with a diagnostic antibody that binds to the same protein as the ADC antibody.

[0024] A third embodiment relates to a method for generating a predictive efficacy score indicative of how a cancer patient will respond to an ADC based on the percentage of cancer cells in the patient's tissue sample that exhibit an optical density of staining that is equal to or greater than an optical density threshold.

[0025] A fourth embodiment relates to identifying cancer patients who would likely benefit from administration of an ADC based on a proximity score equal to the percentage of cancer cells in a patient's tissue sample that have an average optical density of staining above an optical density threshold, or have an average optical density of staining below an optical density threshold but within a predetermined distance from cells whose average optical density of staining is above the optical density threshold.

[0026] A fifth embodiment relates to a method for generating a predictive efficacy score indicative of how a cancer patient will respond to an ADC based on the median absolute deviation (MAD) of the optical density of stained cancer cells in a digital image of the patient's tissue sample from the median optical density of all cancer cells in the digital image, where the predictive efficacy score predicts a positive response of the patient to the ADC if the predictive efficacy score exceeds a predetermined deviation threshold.

[0027] A sixth embodiment relates to a method for generating a predictive efficacy score indicative of how a cancer patient will respond to an ADC based on identifying the 85% quantile of the difference between the optical density of the membrane staining and the optical density of the cytoplasmic staining, the predictive efficacy score predicting a positive response of the patient to the ADC if the predictive efficacy score exceeds a difference threshold.

[0028] I. General definition Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. Singleton, et al., DICTIONARY OF MICROBIOLOGY AND MOLECULAR BIOLOGY, 20 ED., John Wiley and Sons, New York (1994) and Hale & Marham, THE HARPER COLLINS DICTIONARY OF BIOLOGY, Harper Perennial, NY (1991) provide those skilled in the art with a general dictionary of many of the terms used in this disclosure.

[0029] Unless otherwise stated, any nucleic acid sequence is written from left to right in 5' to 3' orientation. Amino acid sequences are written from left to right in amino to carboxy orientation, respectively. Amino acids are referred to herein using the name of the amino acid, its three-letter abbreviation, or its one-letter abbreviation. The term "protein" as used herein includes proteins, polypeptides, and peptides. As used herein, the term "amino acid sequence" is synonymous with the terms "polypeptide" and / or "protein". In some instances, the term "amino acid sequence" is synonymous with the term "peptide". The terms "protein" and "polypeptide" are used interchangeably herein. In the present disclosure and claims, conventional one-letter and three-letter codes for amino acid residues may be used. The three-letter codes for amino acids are defined in accordance with the IUPACIUB Joint Commission on Biochemical Nomenclature (JCBN). Polypeptides may be encoded by multiple nucleotide sequences due to the degeneracy of the genetic code.

[0030] As used herein, the term "about" encompasses an acceptable degree of error for the quantity measured given the nature or precision of the measurement. Exemplary degrees of error are within 20 percent (%), typically within 10%, and more typically within 5% of a given value or range of values. Preferably, the term "about" as used herein includes plus or minus (±) 5% of the numerical value of the number with which it is used, preferably ±4%, ±3%, ±2%, ±1%, ±0.5%, ±0.1%.

[0031] Some embodiments of the novel predictive method disclosed herein involve a quantitative continuous score (QCS). The QCS may indicate a predictive efficacy score, a recommended dosage, or a predictive survival time indicator. As used herein, the term "QCS positive" refers to a cancer that is likely to respond to anti-FRα ADC therapy. The term "QCS negative" refers to a cancer that is unlikely to respond to anti-FRα ADC therapy.

[0032] II. Anti-FRα antibodies in ADCs A. Antibody Sequence The antibody or antigen-binding fragment of the ADC defined herein includes the recited complementarity determining region (CDR) sequences or variable heavy and variable light sequences (reference antibody), as well as functional variants thereof. The functional variant binds to the target antigen in the same manner as the reference antibody, and preferably exhibits antigen cross-reactivity in the same manner as the reference antibody. The functional variant may have a different affinity for the target antigen when compared to the reference antibody, but preferably has substantially the same affinity.

[0033] In some embodiments, functional variants of a reference antibody show sequence variations in one or more CDRs when compared with the corresponding reference CDR sequence. Thus, functional antibody variants can include functional variants of CDRs. When the term "functional variant" is used in the context of a CDR sequence, this means that the CDR has at most two, preferably at most one amino acid difference when compared with the corresponding reference CDR sequence, and when combined with the remaining five CDRs (or variants thereof), allows the variant antibody to bind to the same target antigen as the reference antibody and preferably exhibit the same antigen cross-reactivity as the reference antibody. Functional variants can be referred to as "variant antibodies".

[0034] Figures 1-6 show the CDR, VH and VL, heavy and light chain, FR and constant domain sequences of constructs AB1370049, AB1370026, AB1370035, AB1370083, AB1370095 and AB1370117. In the event of any discrepancy between the sequences listed below and those in Figures 1-6, the sequences in Figures 1-6 take precedence.

[0035] Figure 1 shows various CDRs associated with ADC antibodies. In some embodiments, the ADC antibody or antigen-binding fragment thereof comprises CDRs associated with any one of six constructs AB1370049, AB1370026, AB1370035, AB1370083, AB1370095, or AB1370117, where the CDRs are determined by Kabat, Chothia, or IMGT. Figure 1 lists the sequences of SEQ ID NO:1 to SEQ ID NO:36 of the CDRs defined by Kabat.

[0036] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises the following CDRs:

[0037] Heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1;

[0038] Heavy chain CDR2 of SEQ ID NO:2 (RTYYRSKWYNDYAVSVKS);

[0039] Heavy chain CDR3 of SEQ ID NO:3 (GVGSFDY),

[0040] Light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4,

[0041] Light chain CDR2 (KASGLES) of SEQ ID NO:5, and

[0042] Light chain CDR3 of SEQ ID NO:6 (QQYNSYSQLT).

[0043] FIG. 2 lists the CDR sequences of SEQ ID NO:37 to SEQ ID NO:48.

[0044] In some embodiments, an ADC anti-FRα antibody or antigen-binding fragment thereof has a heavy chain variable region (VH) comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 37, and a light chain variable region (VL) comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, or at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 38. In some embodiments, an ADC anti-FRα antibody or antigen-binding fragment thereof comprises a VH of SEQ ID NO: 37 and a VL of SEQ ID NO:38.

[0045] Figure 3 shows the amino acid sequences of the heavy chain (VH-CH) and light chain (VL-CL) of constructs AB1370049, AB1370026, and AB1370035, as shown in SEQ ID NOs: 49 to 54. Figure 4 shows the amino acid sequences of the VH-CH and VL-CL of constructs AB1370083, AB1370095, and AB1370117, as shown in SEQ ID NOs: 55 to 60.

[0046] In some embodiments, the ADC anti-FRα antibody comprises a heavy chain comprising an amino acid sequence at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 49, and a light chain comprising an amino acid sequence at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 50. In some embodiments, the ADC anti-FRα antibody comprises the heavy chain amino acid sequence of SEQ ID NO: 49 and the light chain amino acid sequence of SEQ ID NO:50.

[0047] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises the following CDRs:

[0048] Heavy chain CDR1 (SYAMS) of SEQ ID NO: 7,

[0049] Heavy chain CDR2 of SEQ ID NO:8 (SISSGRSYIYYADSVKG);

[0050] Heavy chain CDR3 of SEQ ID NO: 9 (EMQQLALDY),

[0051] Light chain CDR1 of SEQ ID NO: 10 (RASQGISNFLA),

[0052] Light chain CDR2 (AASSLQS) of SEQ ID NO: 11, and

[0053] Light chain CDR3 of sequence number 12 (QQYNSYPFT).

[0054] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof has a VH comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 39, and a VL comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 40. In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises a VH of SEQ ID NO: 39 and a VL of SEQ ID NO: 40.

[0055] In some embodiments, the ADC anti-FRα antibody comprises a heavy chain comprising an amino acid sequence at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 51, and a light chain comprising an amino acid sequence at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 52. In some embodiments, the ADC anti-FRα antibody comprises the heavy chain amino acid sequence of SEQ ID NO: 51 and the light chain amino acid sequence of SEQ ID NO:52.

[0056] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises the following CDRs:

[0057] Heavy chain CDR1 of SEQ ID NO: 13 (SNSAAWN),

[0058] Heavy chain CDR2 of SEQ ID NO: 14 (RTYYRSNWYNDYTLSVKS);

[0059] Heavy chain CDR3 of SEQ ID NO: 15 (GVGRFDS),

[0060] Light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 16,

[0061] Light chain CDR2 (KASSLES) of SEQ ID NO: 17, and

[0062] Light chain CDR3 of sequence number 18 (QEYKTYSIFT).

[0063] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof has a VH comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 41, and a VL comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 42. In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises a VH of SEQ ID NO: 41 and a VL of SEQ ID NO: 42.

[0064] In some embodiments, the ADC anti-FRα antibody comprises a heavy chain comprising an amino acid sequence at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 53, and a light chain comprising an amino acid sequence at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 54. In some embodiments, the ADC anti-FRα antibody comprises the heavy chain amino acid sequence of SEQ ID NO: 53 and the light chain amino acid sequence of SEQ ID NO:54.

[0065] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises the following CDRs:

[0066] Heavy chain CDR1 of SEQ ID NO: 19 (SYNMN);

[0067] Heavy chain CDR2 of SEQ ID NO: 20 (SISSGSSYIYYADSMKG);

[0068] Heavy chain CDR3 of SEQ ID NO: 21 (GMTTLTFDY),

[0069] Light chain CDR1 of SEQ ID NO: 22 (RASQGISTFLA),

[0070] Light chain CDR2 (AASSLQS) of SEQ ID NO: 23, and

[0071] Light chain CDR3 of sequence number 24 (QQYISYPLT).

[0072] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof has a VH comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 43, and a VL comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 44. In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises a VH of SEQ ID NO: 43 and a VL of SEQ ID NO: 44.

[0073] In some embodiments, the ADC anti-FRα antibody comprises a heavy chain comprising an amino acid sequence at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 55, and a light chain comprising an amino acid sequence at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 56. In some embodiments, the ADC anti-FRα antibody comprises the heavy chain amino acid sequence of SEQ ID NO: 55 and the light chain amino acid sequence of SEQ ID NO:56.

[0074] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises the following CDRs:

[0075] Heavy chain CDR1 (SYSMN) of SEQ ID NO: 25;

[0076] Heavy chain CDR2 of SEQ ID NO: 26 (SISSRSSYVYYADSVKG),

[0077] Heavy chain CDR3 of SEQ ID NO: 27 (GMTTLTFDY),

[0078] Light chain CDR1 of SEQ ID NO: 28 (RASQGISSFLA),

[0079] Light chain CDR2 (AASSLQS) of SEQ ID NO: 29, and

[0080] Light chain CDR3 of sequence number 30 (QQYNSYPLT).

[0081] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof has a VH comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 45, and a VL comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 46. In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises a VH of SEQ ID NO: 45 and a VL of SEQ ID NO:46.

[0082] In some embodiments, the ADC anti-FRα antibody comprises a heavy chain comprising an amino acid sequence at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 57, and a light chain comprising an amino acid sequence at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 58. In some embodiments, the ADC anti-FRα antibody comprises the heavy chain amino acid sequence of SEQ ID NO: 57 and the light chain amino acid sequence of SEQ ID NO:58.

[0083] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises the following CDRs:

[0084] Heavy chain CDR1 of SEQ ID NO: 31 (SDSATWN),

[0085] Heavy chain CDR2 of SEQ ID NO: 32 (RTYYRSKWYSDYAVSVKS);

[0086] Heavy chain CDR3 of SEQ ID NO: 33 (GGAPFDY),

[0087] Light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 34,

[0088] Light chain CDR2 (KASSLES) of SEQ ID NO: 35, and

[0089] Light chain CDR3 of sequence number 36 (QQYNSYSMYT).

[0090] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof has a VH comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 47, and a VL comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 48. In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises a VH of SEQ ID NO: 47 and a VL of SEQ ID NO: 48.

[0091] In some embodiments, the ADC anti-FRα antibody comprises a heavy chain comprising an amino acid sequence at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 59, and a light chain comprising an amino acid sequence at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 60. In some embodiments, the ADC anti-FRα antibody comprises the heavy chain amino acid sequence of SEQ ID NO: 59 and the light chain amino acid sequence of SEQ ID NO: 60.

[0092] FIG. 5 shows the amino acid sequences of SEQ ID NO: 61 to SEQ ID NO: 108 of the heavy and light chain framework regions of constructs AB1370049, AB1370026, AB1370035, AB1370083, AB1370095, and AB1370117.

[0093] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises: (a) a light chain VL-FR1, VL-FR2, VL-FR3, and VL-FR4 that are at least 80%, 85%, 90%, or 95% identical, or identical to reference light chain VL-FR1, VL-FR2, VL-FR3, and VL-FR4, respectively, of any one of constructs AB1370049, AB1370026, AB1370035, AB1370083, AB1370095, or AB1370117. 5, (a) VL-FR3, and VL-FR4, and (b) heavy chain VH-FR1, VH-FR2, VH-FR3, and VH-FR4 that are at least 80%, 85%, 90%, or 95% identical, or identical to the reference heavy chain VH-FR1, VH-FR2, VH-FR3, and VH-FR4, respectively, of any one of constructs AB1370049, AB1370026, AB1370035, AB1370083, AB1370095, or AB1370117 shown in FIG. 5.

[0094] FIG. 6 lists the amino acid sequences of SEQ ID NO: 109 to SEQ ID NO: 111 of the heavy chain constant domain, the light chain constant domain, and the heavy chain constant domain CH1.

[0095] In some embodiments, an ADC anti-FRα antibody comprises a constant heavy chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 109, and a constant light chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 110. In some embodiments, an ADC anti-FRα antibody comprises a constant heavy chain amino acid sequence of SEQ ID NO: 109 and a constant light chain amino acid sequence of SEQ ID NO:110.

[0096] In some embodiments, an ADC anti-FRα antigen-binding fragment comprises a constant heavy chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 111, and a constant light chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 110. In some embodiments, an ADC anti-FRα antigen-binding fragment comprises the constant heavy chain amino acid sequence of SEQ ID NO: 111 and the constant light chain amino acid sequence of SEQ ID NO:110.

[0097] ADC antibodies may include slight variations in amino acid sequence, provided that the variations in amino acid sequence maintain at least 75%, more preferably at least 80%, at least 90%, at least 95%, and most preferably at least 99% sequence identity with an ADC antibody or antigen-binding fragment thereof defined anywhere herein.

[0098] ADC antibodies may include variants in which an amino acid residue from one species is substituted with the corresponding residue from another species, either at a conserved or non-conserved position. In some embodiments, an amino acid residue at a non-conserved position is substituted with a conserved or non-conserved residue. Specifically, conservative amino acid substitutions are contemplated.

[0099] "Conservative amino acid substitution" refers to an amino acid residue that is replaced with an amino acid residue that has a similar side chain. Families of amino acid residues with similar side chains have been defined in the art and include basic side chains (e.g., lysine, arginine, or histidine), acidic side chains (e.g., aspartic acid or glutamic acid), uncharged polar side chains (e.g., glycine, asparagine, glutamine, serine, threonine, tyrosine, or cysteine), non-polar side chains (e.g., alanine, valine, leucine, isoleucine, proline, phenylalanine, methionine, or tryptophan), β-branched side chains (e.g., threonine, valine, isoleucine), and aromatic side chains (e.g., tyrosine, phenylalanine, tryptophan, or histidine). Thus, when an amino acid in a polypeptide is replaced with another amino acid from the same side chain family, the amino acid substitution is considered to be conservative. The inclusion of conservatively modified variants in the ADC antibodies does not exclude other forms of variants, for example, polymorphic variants, interspecies homologs, and alleles.

[0100] "Non-conservative amino acid substitutions" include those in which (i) a residue having an electropositive side chain (e.g., Arg, His, or Lys) is replaced with or by an electronegative residue (e.g., Glu or Asp); (ii) a hydrophilic residue (e.g., Ser or Thr) is replaced with or by a hydrophobic residue (e.g., Ala, Leu, Ile, Phe, or Val); (iii) a cysteine ​​or proline is replaced with or by any other residue; or (iv) a residue having a bulky hydrophobic or aromatic side chain (e.g., Val, His, Ile, or Trp) is replaced with or by a residue having a smaller side chain (e.g., Ala or Ser) or no side chain (e.g., Gly).

[0101] In addition to the 20 standard amino acids, non-standard amino acids (e.g., 4-hydroxyproline, 6-N-methyllysine, 2-aminoisobutyric acid, isovaline, and α-methylserine) may be substituted for amino acid residues in the ADC antibodies. A limited number of non-conserved amino acids, amino acids that are not encoded by the genetic code, and unnatural amino acids may be substituted for amino acid residues. The ADC antibodies may also include non-naturally occurring amino acid residues.

[0102] Non-naturally occurring amino acids include, but are not limited to, trans-3-methylproline, 2,4-methanoproline, cis-4-hydroxyproline, trans-4-hydroxy-proline, N-methylglycine, allo-threonine, methylthreonine, hydroxy-ethylcysteine, hydroxyethylhomo-cysteine, nitro-glutamine, homoglutamine, pipecolic acid, tert-leucine, norvaline, 2-azaphenylalanine, 3-azaphenyl-alanine, 4-azaphenyl-alanine, and 4-fluorophenylalanine. Several methods are known in the art for incorporating non-naturally occurring amino acid residues into proteins. For example, an in vitro system can be used in which nonsense mutations are suppressed using chemically aminoacylated suppressor tRNAs. Methods for synthesizing amino acids and aminoacylated tRNAs are known in the art. Transcription and translation of plasmids containing nonsense mutations are performed in a cell-free system that includes E. coli S30 extracts and commercially available enzymes and other reagents. Proteins are purified by chromatography (see, e.g., Robertson et al., J. Am. Chem. Soc. 113:2722, 1991; Ellman et al., Methods Enzymol. 202:301, 1991; Chung et al., Science 259:806-9, 1993; and Chung et al., Proc. Natl. Acad. Sci. USA 90:10145-9, 1993). In the second method, translation is carried out in Xenopus oocytes by microinjection of mutant mRNA and chemically aminoacylated suppressor tRNA (Turcatti et al., J. Biol. Chem. 271:1991-8, 1996). In a third method, E. coli cells are cultured in the absence of the natural amino acid to be substituted (e.g., phenylalanine) and in the presence of the desired unnatural amino acid(s) (e.g., 2-azaphenylalanine, 3-azaphenylalanine, 4-azaphenylalanine, or 4-fluorophenylalanine).Non-naturally occurring amino acids are incorporated into polypeptides in place of their natural counterparts. See Koide et al., Biochem. 33:7470-6, 1994. Naturally occurring amino acid residues can be converted to non-naturally occurring species by in vitro chemical modification. Chemical modification can be combined with site-directed mutagenesis to further expand the scope of substitutions (Wynn and Richards, Protein Sci. 2:395-403, 1993).

[0103] A limited number of non-conserved amino acids, amino acids that are not encoded by the genetic code, non-naturally occurring amino acids, and unnatural amino acids may be substituted for amino acid residues of the ADC antibody.

[0104] Essential amino acids in ADC antibodies can be identified according to procedures known in the art, such as site-directed mutagenesis or alanine scanning mutagenesis (Cunningham and Wells, Science 244:1081-5, 1989). Sites of biological interaction can also be determined by physical analysis of the structure, as determined by techniques such as nuclear magnetic resonance, crystallography, electron diffraction or photoaffinity labeling, in conjunction with mutations of putative contact site amino acids. See, for example, de Vos et al., Science 255:306-12, 1992; Smith et al., J. Mol. Biol. 224:899-904, 1992; Wlodaver et al., FEBS Lett. 309:59-64, 1992. The identity of essential amino acids can also be inferred from analysis of homology with related components of ADC antibodies (e.g., translocation or protease components).

[0105] Multiple amino acid substitutions can be made and tested using known methods of mutagenesis and screening, such as those disclosed by Reidhaar-Olson and Sauer (Science 241:53-7, 1988) or Bowie and Sauer (Proc. Natl. Acad. Sci. USA 86:2152-6, 1989). These authors disclose methods in which two or more positions in a polypeptide are randomized simultaneously, functional polypeptides are selected, and the mutagenized polypeptides are then sequenced to determine the range of substitutions tolerated at each position. Other methods that can be used include phage display (e.g., Lowman et al., Biochem. 30:10832-7, 1991; Ladner et al., U.S. Pat. No. 5,223,409; Huse, WIPO Publication No. WO 92 / 06204) and region-directed mutagenesis (Derbyshire et al., Gene 46:145, 1986; Ner et al., DNA 7:127, 1988).

[0106] The "percent sequence identity" between two or more nucleic acid or amino acid sequences is a function of the number of identical positions shared by those sequences.Thus, the percent sequence identity can be calculated by dividing the number of identical nucleotides / amino acids by the total number of nucleotides / amino acids and multiplying by 100.The calculation of percent sequence identity can also take into account the number of gaps and the length of each gap that need to be introduced to optimize the alignment of two or more sequences.The sequence comparison and identity percent determination between two or more sequences can be performed using a specific mathematical algorithm such as BLAST, which is well known to those skilled in the art.

[0107] Any of a variety of sequence alignment methods can be used to determine percent sequence identity, including but not limited to global, local, and hybrid (e.g., segment approach) methods. Protocols for determining percent sequence identity are routine procedures within the scope of those skilled in the art. Global methods align sequences from the beginning to the end of a molecule, sum up the scores of individual residue pairs, and apply gap penalties to determine the best alignment. Non-limiting methods include, for example, CLUSTAL W. (see, e.g., Julie D. Thompson et al., CLUSTAL W: Improving the Sensitivity of Progressive Multiple Sequence Alignment Through Sequence Weighting, Position-Specific Gap Penalties and Weight Matrix Choice, 22(22) Nucleic Acids Research 4673-4680(1994)) and iterative refinement (see, e.g., Osamu Gotoh, Significant Improvement in Accuracy of Multiple Protein Sequence Alignments by Iterative Refinement as Assessed by Reference to Structural Alignments, 264(4) J.MoI.Biol.823-838(1996)). Local methods align sequences by identifying one or more conserved motifs shared by all of the input sequences.Non-limiting methods include, for example, Match-box (see, e.g., Eric Depiereux and Ernest Feytmans, Match-Box: A Fundamentally New Algorithm for the Simultaneous Alignment of Several Protein Sequences, 8(5) CABIOS 501-509(1992)), Gibbs sampling (see, e.g., CE Lawrence et al., Detecting Subtle Sequence Signals: A Gibbs Sampling Strategy for Multiple Alignment, 262(5131) Science 208-214(1993)), and Align-M (see, e.g., Ivo Van WaIIe et al., Align-M: A New Algorithm for Multiple Alignment of Highly Divergent Sequences, 20(9) Bioinformatics:1428-1435(2004)).

[0108] Percent sequence identity can be determined by conventional methods. See, e.g., Altschul et al. Bull. Math. Bio. 48:603-16, 1986 and Henikoff and Henikoff, Proc. Natl. Acad. Sci. USA 89:10915-19, 1992. The two amino acid sequences are aligned and the alignment score is optimized using a gap opening penalty of 10, a gap extension penalty of 1, and the "blosum 62" scoring matrix of Henikoff and Henikoff (ibid.).

[0109] In some embodiments, the variable domains of both the heavy and light chains of the ADC antibody or antigen-binding fragment thereof are modified by at least partial replacement of one or more CDRs and / or by replacement and sequence changes of partial framework regions. The CDRs may be derived from antibodies of the same class or even subclass as the antibody from which the framework regions are derived, but it is envisioned that the CDRs are derived from antibodies of different classes and, in certain embodiments, from antibodies from different species. It is not necessary to replace all CDRs with the complete CDRs from the donor variable region to transfer the antigen-binding capacity of one variable domain to another. Rather, it is necessary to transfer only those residues that are necessary to maintain the activity of the antigen-binding site. Given the descriptions provided in U.S. Pat. Nos. 5,585,089, 5,693,761, and 5,693,762 (each of which is incorporated herein by reference), it will be well within the capabilities of one of ordinary skill in the art to carry out routine experimentation to obtain functional antibodies with reduced immunogenicity.

[0110] In some embodiments, the ADC antibody or antigen-binding fragment thereof can include a heavy chain constant region or fragment thereof in addition to a VH and a VL. In some embodiments, the heavy chain constant region is a human heavy chain constant region, e.g., a human IgG constant region or a human IgG1 constant region.

[0111] In some embodiments, residues are inserted into the heavy chain constant region for site-specific conjugation. For example, a cysteine ​​residue may be inserted between amino acids S239 and V240 in the CH2 region of IgG1, which may be referred to as "239 insertion" or "239i".

[0112] In some embodiments, the ADC antibodies disclosed herein may be modified to include alterations or modifications to one or more of the three heavy chain constant domains (CH1, CH2, or CH3) and / or light chain constant domains (CL). In some embodiments, modified constant regions in which one or more domains are partially or completely deleted are contemplated. In some embodiments, the modified ADC antibodies include domain deleted constructs or variants in which the entire CH2 domain has been removed (ΔCH2 constructs). In some embodiments, the omitted constant region domains may be replaced by short amino acid spacers (e.g., 10 residues) that provide some of the molecular flexibility typically conferred by deleted constant regions. Deletion or inactivation (by point mutation or other means) of constant region domains can reduce Fc receptor binding of circulating modified antibodies. In other cases, constant region modifications can mitigate complement binding and thus reduce serum half-life and nonspecific association of conjugated ADC payloads (e.g., cytotoxins). Still other modifications of the constant region can be used to remove disulfide bonds or oligosaccharide moieties that allow for enhanced localization due to increased antigen specificity or antibody flexibility. In some embodiments, an ADC antibody or antigen-binding fragment thereof does not have antibody-dependent cellular cytotoxicity (ADCC) activity and / or complement-dependent cytotoxicity (CDC) activity.

[0113] In some embodiments, the ADC antibodies or antigen-binding fragments thereof can be engineered to fuse the CH3 domain directly to the hinge region of the respective modified antibody or fragment thereof. In other constructs, a peptide spacer can be inserted between the hinge region and the modified CH2 and / or CH3 domains. For example, a compatible construct can be expressed in which the CH2 domain is deleted and the remaining CH3 domain (modified or unmodified) is linked to the hinge region with a spacer of 5-20 amino acids. Such a spacer can be added, for example, to ensure that the regulatory elements of the constant domain are free and accessible or that the hinge region is flexible. Amino acid spacers can, in some cases, prove immunogenic and induce undesirable immune responses against the construct. In some embodiments, to maintain the desired biochemical qualities of the modified antibody, any spacer added to the construct can be relatively non-immunogenic or even omitted entirely.

[0114] In addition to the deletion of entire constant region domains, the ADC antibodies or antigen-binding fragments thereof provided herein may be modified by partial deletion or substitution of several or even single amino acids in the constant region. For example, mutation of a single amino acid in selected regions of the CH2 domain may be sufficient to reduce Fc binding and thereby increase tumor localization. Similarly, one or more constant region domains that control effector functions (e.g., complement C1Q binding) can be deleted in whole or in part. Such partial deletion of the constant region can improve selected characteristics of the ADC antibody or antigen-binding fragment thereof (e.g., serum half-life) while leaving other desirable functions associated with the constant region domain of interest intact. Furthermore, the constant regions of the ADC antibodies and antigen-binding fragments thereof can be modified by mutation or substitution of one or more amino acids that enhance the profile of the resulting construct. In this regard, it is possible to disrupt the activity provided by a conserved binding site (e.g., Fc binding) while substantially maintaining the configuration and immunogenicity profile of the modified antibody or antigen-binding fragment thereof. In some embodiments, one or more amino acids can be added to the constant region to enhance a desirable characteristic, such as reducing or increasing effector function, or to provide more ADC payload (e.g., cytotoxin) or carbohydrate binding. In some embodiments, it may be desirable to insert or duplicate a particular sequence from a selected constant region domain.

[0115] In some embodiments, the heavy chain constant region or fragment thereof (e.g., a human IgG constant region or fragment thereof) can contain one or more amino acid substitutions relative to a wild-type IgG constant domain, such that the modified IgG has an increased half-life compared to the half-life of an IgG having a wild-type IgG constant domain. For example, the IgG constant domain can contain one or more amino acid substitutions at amino acid residues at positions 251-257, 285-290, 308-314, 385-389, and 428-436 (wherein the amino acid position numbering is according to the EU index as set forth in Kabat). In some embodiments, the IgG constant domain comprises a substitution of the amino acid at Kabat position 252 with tyrosine (Y), phenylalanine (F), tryptophan (W), or threonine (T); a substitution of the amino acid at Kabat position 254 with threonine (T); a substitution of the amino acid at Kabat position 256 with serine (S), arginine (R), glutamine (Q), glutamic acid (E), aspartic acid (D), or threonine (T); a substitution of the amino acid at Kabat position 257 with leucine (L); a substitution of the amino acid at Kabat position 309 with proline (P). substitution of the amino acid at Kabat position 311 with serine (S); substitution of the amino acid at Kabat position 428 with threonine (T), leucine (L), phenylalanine (F), or serine (S); substitution of the amino acid at Kabat position 433 with arginine (R), serine (S), isoleucine (I), proline (P), or glutamine (Q); or substitution of the amino acid at Kabat position 434 with tryptophan (W), methionine (M), serine (S), histidine (H), phenylalanine (F), or tyrosine.

[0116] In some embodiments, the ADC antibody or antigen-binding fragment thereof comprises a YTE variant. The term "YTE" or "YTE variant" refers to a mutation in IgG1 Fc that results in increased binding to human FcRn and improves the serum half-life of the antibody carrying the mutation. The YTE variant comprises a combination of three mutations, M252Y / S254T / T256E (EU numbering Kabat et al. (1991) Sequences of Proteins of Immunological Interest, USPublic Health Service, National Institutes of Health, Washington, DC), introduced into the heavy chain of IgG1. See U.S. Patent No. 7,658,921 (incorporated herein by reference). The YTE variant has been shown to increase the serum half-life of an antibody by approximately 4-fold compared to the wild-type version of the same antibody (Dall'Acqua et al., J. Biol. Chem. 281:23514-24 (2006); Robbie et al., (2013) Antimicrob. Agents Chemother. 57, 6147-6153). See also U.S. Patent No. 7,083,784, which is incorporated herein by reference in its entirety.

[0117] In some embodiments, the ADC antibody or antigen-binding fragment thereof comprises:

[0118] L at the N-terminus of VH (e.g., position 1);

[0119] an E at the N-terminus (e.g., position 1) of VH, or

[0120] Q at the N-terminus of VH (e.g., position 1).

[0121] B. Definitions and Antibody Types As used herein, the term "antibody" refers to an immunoglobulin molecule that specifically binds to or is immunologically reactive with a particular antigen.

[0122] Generally, an antibody comprises at least two "light chains" (LC) and two "heavy chains" (HC). The light and heavy chains of such antibodies are polypeptides comprising several domains. Each heavy chain comprises a heavy chain variable region (abbreviated herein as "VH") and a heavy chain constant region (abbreviated herein as "CH"). The heavy chain constant region comprises heavy chain constant domains CH1, CH2, and CH3 (antibody classes IgA, IgD, and IgG), and optionally heavy chain constant domain CH4 (antibody classes IgE and IgM). Each light chain comprises a light chain variable domain (abbreviated herein as "VL") and a light chain constant domain (abbreviated herein as "CL").

[0123] In some embodiments, the ADC antibody is a full length antibody. An "intact" or "full length" antibody, as used herein, refers to an antibody having two heavy (H) chain polypeptides and two light (L) chain polypeptides interconnected by disulfide bonds.

[0124] The "variable region" of an antibody refers to the variable region of an antibody light chain or the variable region of an antibody heavy chain, either alone or in combination. The variable regions VH and VL can be further subdivided into hypervariable regions called complementarity determining regions (CDRs) (also known as hypervariable regions), interspersed with more conserved regions called framework regions (FRs). Preferably, each VH and VL is composed of three CDRs and four FRs arranged from amino-terminus to carboxy-terminus in the following order: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4. The VH or VL chain of an antibody may further comprise all or a portion of a heavy chain constant region or a light chain constant region.

[0125] The binding between an antibody and its target antigen or epitope is mediated by CDR. The term "epitope" refers to a target protein region (e.g., polypeptide) that can bind (e.g., be bound by) an antibody or antigen-binding fragment. CDR is the main determinant of antigen specificity. There are at least two techniques for determining CDR: (1) an approach based on cross-species sequence variability (i.e., Kabat et al. Sequences of Proteins of Immunological Interest, (5th ed., 1991, National Institutes of Health, Bethesda Md.)), and (2) an approach based on crystallographic studies of antigen-antibody complexes (Al-lazikani et al. (1997) J. Molec. Biol. 273:927-948)). In addition, the art may use a combination of these two methods to determine CDR.

[0126] The sequence of CDR can be identified by referring to any numbering system known in the art. Figure 7 shows three exemplary numbering systems: the Kabat system (Kabat, EA, et al., Sequences of Proteins of Immunological Interest, 5th ed., Public Health Service, National Institutes of Health, Bethesda, MD (1991)), the Chothia system (Chothia &,Lesk, "Canonical Structures for the Hypervariable Regions of Immunoglobulins," J. Mol. Biol. 196, 901-917 (1987)), and the IMGT system (Lefranc et al., "IMGT Unique Numbering for Immunoglobulin and Cell Receptor Variable Domains and Ig superfamily V-like domains," Dev. Comp. Immunol. 27, 55-77 (2003)).

[0127] The "constant domains" (or "constant regions") of the heavy and light chains are not directly involved in binding the antibody to its target, but instead exhibit various effector functions. The constant regions of the antibody can mediate the binding of the immunoglobulin to host tissues or factors, including various cells of the immune system (e.g., effector cells) and the first component (Clq) of the classical complement system.

[0128] There are five major classes of heavy chain constant regions, classified as IgA, IgG, IgD, IgE, and IgM, each with a characteristic effector function designated by the isotype. Ig molecules interact with multiple classes of cell receptors. For example, IgG molecules interact with three classes of Fcγ receptors (FcγRs) specific for IgG class antibodies: FcγRI, FcγRII, and FcγRIII. Binding of antibodies to Fc receptors on the cell surface elicits several important and diverse biological responses, including phagocytosis and destruction of antibody-coated particles, clearance of immune complexes, lysis of antibody-coated target cells by killer cells (called antibody-dependent cell-mediated cytotoxicity, or ADCC), release of inflammatory mediators, placental transfer, and control of immunoglobulin production. It has been reported that the critical sequences for binding of IgG to FcγR receptors are located in the CH2 and CH3 domains.

[0129] In a preferred embodiment, the ADC anti-FRα antibody or antigen-binding fragment thereof is an IgG isotype. The ADC anti-FRα antibody or antigen-binding fragment thereof can be any IgG subclass, for example, IgG1, IgG2, IgG3, or IgG4 isotype. In a preferred embodiment, the ADC anti-FRα antibody or antigen-binding fragment thereof is based on the IgG1 isotype. The use of wild-type human IgG1 molecules, which are closer to natural IgG, can reduce pathogenesis and other risks.

[0130] For the amino acid positions of the heavy chain constant region of the ADC antibodies described herein, the numbering is according to the EU index as originally described in Edelman, GM, et al., Proc. Natl. Acad. Sci. USA 63 (1969) 78-85. Edelman's EU numbering is also described in Kabat et al. (1991) (supra). Thus, the term "EU index according to Kabat", "EU index", "Kabat's EU index", or "EU numbering" in the context of heavy chains refers to the residue numbering system based on the human IgG1 EU antibody of Edelman et al. as described in Kabat et al. (1991). The numbering system used for the light chain constant region amino acid sequence is similarly described in Kabat et al. (supra). Thus, as used herein, "numbered according to Kabat" refers to the Kabat numbering system as described in Kabat et al. (supra).

[0131] The terms "Fc region", "Fc portion" and "Fc" are used interchangeably herein and refer to the portion of a native immunoglobulin formed by two Fc chains. Each "Fc chain" comprises a constant domain CH2 and a constant domain CH3. Each Fc chain may also comprise a hinge region. A native Fc region is a homodimer. In some embodiments, an Fc region may comprise modifications to force Fc heterodimerization and therefore may be a heterodimer. The Fc region comprises carbohydrate moieties and binding sites for complement and Fc receptors (including the FcRn receptor) and does not have antigen binding activity. Fc can refer to this region alone or in the context of an antibody, antibody fragment, or Fc fusion protein. Polymorphisms have been found at several Fc domain sites, including but not limited to EU positions 270, 272, 312, 315, 356, and 358, resulting in slight variations between the sequences described in this application and those known in the art. Consequently, all naturally occurring IgG Fc regions are referred to as "wild-type IgG Fc domains" or "WT IgG Fc domains" (i.e., any allele). Human IgG1, IgG2, IgG3, and IgG4 heavy chain sequences can be obtained in various sequence databases, including the UniProt database (www.uniprot.org) (under accession numbers P01857 (IGHG1_HUMAN), P01859 (IGHG2_HUMAN), P01860 (IGHG3_HUMAN), and P01861 (IGHG4_HUMAN), respectively).

[0132] In some embodiments, the ADC anti-FRα antibody is a monoclonal antibody. A "monoclonal antibody" (mAb) refers to a homogenous antibody population that is involved in highly specific recognition and binding of a single antigenic determinant or epitope. This is in contrast to a polyclonal antibody, which usually contains different antibodies against different antigenic determinants. The term "monoclonal antibody" can encompass both full-length monoclonal antibodies, as well as antibody fragments (e.g., Fab, Fab', F(ab')2, Fv), single-chain (scFv) variants, fusion proteins containing antibody moieties, and any other modified immunoglobulin molecule that contains an antigen recognition site. Furthermore, "monoclonal antibody" refers to such antibodies made in any number of ways, including but not limited to hybridoma, phage selection, recombinant expression, and transgenic animals. More preferably, the ADC anti-FRα antibody is an isolated monoclonal antibody. In a more preferred embodiment, the ADC antibody is a fully human monoclonal antibody.

[0133] ADC anti-FRα antibodies and their antigen-binding fragments can be derived from any species by recombinant means. For example, ADC antibodies or antigen-binding fragments can be mouse, rat, goat, horse, pig, cow, chicken, rabbit, camel, donkey, human, or chimeric versions thereof. For use in human administration, non-human derived ADC antibodies or antigen-binding fragments can be genetically or structurally modified to be less immunogenic when administered to human patients. Human or humanized antibodies are particularly preferred, especially as recombinant human or humanized antibodies.

[0134] The term "human antibody" refers to an antibody produced in a human or an antibody having an amino acid sequence corresponding to an antibody produced in a human produced using any technique known in the art. A human antibody may include an intact or full-length antibody, a fragment thereof, and / or an antibody comprising at least one human heavy and / or light chain polypeptide (e.g., an antibody comprising a mouse light chain and a human heavy chain polypeptide). A human antibody may include amino acid residues not encoded by human germline immunoglobulin sequences (e.g., mutations introduced by random or site-specific mutagenesis in vitro, or during gene rearrangement or by somatic mutation in vivo). A human antibody may be produced in human cells (through recombinant expression), non-human animals, or in prokaryotic or eukaryotic cells capable of expressing functionally rearranged human immunoglobulin (such as heavy and light) genes. A linker peptide not found in naturally occurring human antibodies may be included in a single-chain human antibody. For example, an Fv can have a linker peptide, such as two to about eight glycine or other amino acid residues, linking the variable region of the heavy chain and the variable region of the light chain. These linker peptides are considered to be of human origin. Human antibodies can be produced using a variety of techniques, including phage display techniques using antibody libraries derived from human immunoglobulin sequences. Transgenic mice incapable of expressing functional native immunoglobulins, but which can express human immunoglobulin genes, can also be used to make human antibodies (see, e.g., PCT Publication Nos. WO1998 / 24893, WO1992 / 01047, WO1996 / 34096, WO1996 / 33735, U.S. Pat. No. 5,213,133). (See, e.g., 5,413,923, 5,625,126, 5,633,425, 5,569,825, 5,661,016, 5,545,806, 5,814,318, 5,885,793, 5,916,771, and 5,939,598, each of which is incorporated herein by reference.) Human antibodies can also be directly prepared using a variety of techniques known in the art.Immortalized human B lymphocytes produced in vitro or isolated from immunized individuals can be generated that produce antibodies against target antigens.See, for example, Cole et al., Monoclonal Antibodies and Cancer Therapy, Alan R. Liss, p.77 (1985); Boemer et al., J. Immunol. 147(1):86-95 (1991); U.S. Patent No. 5,750,373.

[0135] The term "humanized antibody" refers to an antibody in which the framework or CDRs have been modified to include CDRs of an immunoglobulin of different specificity compared to the parent immunoglobulin. For example, murine CDRs can be grafted into the framework regions of a human antibody to prepare a "humanized antibody." See, for example, Riechmann, L., et al., Nature 332 (1988) 323-327, and Neuberger, MS, et al., Nature 314 (1985) 268-270. In some embodiments, a "humanized antibody" is one in which the constant regions have been further modified or altered from those of the original antibody to generate desired properties.

[0136] Humanized antibodies can optionally be prepared by a process of analysis of parental sequences and various conceptual humanized and engineered products using three-dimensional models of parental, engineered, and humanized sequences. Three-dimensional immunoglobulin models are commonly available and are familiar to those skilled in the art. Computer programs are available that illustrate and display possible three-dimensional conformations of selected candidate immunoglobulin sequences. Inspection of these displays allows analysis of the likely role of the residues in the functioning of the candidate immunoglobulin sequence, i.e., analysis of residues that influence the ability of the candidate immunoglobulin to bind to an antigen, such as FRα. In this way, folate receptor (FR) residues can be selected and combined with consensus and import sequences such that desired antibody properties, such as increased affinity for the target antigen(s), are achieved.

[0137] Humanized antibodies can be further modified by substitution of additional residues either within the Fv framework regions and / or within the substituted non-human residues to refine and optimize the specificity, affinity, and / or potency of the antibody. Generally, a humanized antibody contains substantially all of at least one, typically two or three, variable domains, including all or substantially all of the CDR regions corresponding to a non-human immunoglobulin, while all or substantially all of the FR regions are of a human immunoglobulin consensus sequence. A humanized antibody may also contain at least a portion of an immunoglobulin constant region or domain (Fc), typically that of a human immunoglobulin. Examples of methods used to generate humanized antibodies are described in U.S. Patent Nos. 5,225,539 and 5,639,641, each of which is incorporated herein by reference.

[0138] The term "chimeric antibody" refers to an antibody that contains a variable region (i.e., a binding region) from one source or species and at least a portion of a constant region from a different source or species, usually prepared by recombinant DNA technology. Chimeric antibodies that contain a murine variable region and a human constant region are preferred. Other preferred forms of "chimeric antibodies" are those in which the constant region has been modified or changed from the constant region of the original antibody to generate desired properties. Such chimeric antibodies are also referred to as "class-switched antibodies." Chimeric antibodies are the product of expressed immunoglobulin genes that contain DNA segments encoding immunoglobulin variable regions and DNA segments encoding immunoglobulin constant regions. Methods for producing chimeric antibodies, including conventional recombinant DNA and gene transfection techniques, are well known in the art. See, e.g., Morrison, SL, et al., Proc. Natl. Acad. Sci. USA 81 (1984) 6851-6855, U.S. Pat. Nos. 5,202,238 and 5,204,244, each of which is incorporated herein by reference.

[0139] In some embodiments, the ADC antibody is a full-length antibody as described above. Alternatively, the ADC antibody can be an antigen-binding fragment. As used herein, the term "antigen-binding fragment" includes any naturally occurring or artificially constructed configuration of an antigen-binding polypeptide that contains one, two, or three light chain CDRs and / or one, two, or three heavy chain CDRs, and the polypeptide is capable of binding to an antigen.

[0140] In some embodiments, the ADC antigen-binding fragment is a Fab fragment. The ADC antibody can also be a Fab', Fv, scFv, Fd, VNAR domain, IgNAR, intrabody, IgG CH2, minibody, single domain antibody, Fcab, scFv-Fc, F(ab')2, di-scFv, bispecific T cell engager (BiTE®), F(ab')3, tetrabody, triabody, diabody, DVD-Ig, (scFv)2, mAb2, or DARPin.

[0141] The terms "Fab fragment" and "Fab" are used interchangeably herein and comprise a single light chain (e.g., constant domains CL and VL) and a single heavy chain (e.g., constant domains CH1 and VH). The heavy chain of a Fab fragment cannot form disulfide bonds with another heavy chain.

[0142] A "Fab' fragment" contains a single light chain and a single heavy chain, but in addition to CH1 and VH, the "Fab' fragment" also contains the region of the heavy chain between the CH1 and CH2 domains necessary for interchain disulfide bond formation. Thus, two "Fab' fragments" can associate through disulfide bond formation to form a F(ab')2 molecule.

[0143] A "F(ab')2 fragment" contains two light chains and two heavy chains, each of which contains a portion of the constant region necessary for interchain disulfide bond formation between the two heavy chains.

[0144] An "Fv fragment" contains only the variable regions of the heavy and light chains. It does not contain the constant regions.

[0145] A "single domain antibody" is an antibody fragment comprising a single antibody domain unit (eg VH or VL).

[0146] A "single-chain Fv" ("scFv") is an antibody fragment that comprises the VH and VL domains of an antibody linked together to form a single chain. A polypeptide linker is typically used to connect the VH and VL domains of an scFv.

[0147] "Tandem scFv," also known as TandAb®, is a single-chain Fv molecule formed by the covalent association of two scFvs in a tandem orientation with a flexible peptide linker.

[0148] "Bispecific T cell engagers" (BiTEs®) are fusion proteins that contain two single-chain variable fragments (scFvs) on a single peptide chain. One of the scFvs binds to T cells via the CD3 receptor, and the other binds to a tumor cell antigen.

[0149] "Diabodies" are small bivalent bispecific antibody fragments that contain a heavy chain variable domain (VH) connected to a light chain variable domain (VL) on the same polypeptide chain (VH-VL) by a peptide linker that is too short to allow pairing between the two domains on the same chain (Kipriyanov, Int. J. Cancer 77 (1998), 763-772). This allows pairing with the complementary domains of another chain, promoting the assembly of a dimeric molecule with two functional antigen-binding sites.

[0150] "DARPins" are bispecific ankyrin repeat molecules. DARPins are derived from the natural ankyrin protein, which can be found in the human genome and is one of the most abundant types of binding proteins. DARPin library modules are defined by the natural ankyrin repeat protein sequence, with 229 ankyrin repeats used in the initial design and 2200 more ankyrin repeats in subsequent refinements. Modules serve as building blocks for DARPin libraries. Library modules resemble human genome sequences. DARPins are composed of 4-6 modules. Each module is approximately 3.5 kDa, so the size of an average DARPin is 16-21 kDa. Selection of binders is performed by ribosome display, which is completely cell-free and described in He M.and Taussig MJ.,Biochem Soc Trans.2007,Nov;35(Pt 5):962-5.

[0151] In some embodiments, the ADC antibody or antigen-binding fragment thereof may be further modified to include additional chemical moieties that are not normally part of the protein. These derivatized moieties may improve the solubility, biological half-life, or absorption of the protein. The moieties may also reduce or eliminate any undesirable side effects of the protein, etc. A summary of these moieties can be found in Remington's Pharmaceutical Sciences, 22nd ed., Ed. Lloyd V. Allen, Jr. (2012).

[0152] C.FRα binding In preferred embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof specifically binds to FRα. The term "specifically binds to FRα" refers to an antibody that can bind to a defined target with sufficient affinity such that the antibody is useful as a therapeutic agent in targeting FRα. In some embodiments, an ADC antibody that specifically binds to FRα does not bind to other antigens or does not bind to other antigens with sufficient affinity to produce a physiological effect. In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof specifically binds to human FRα (UniProt ID: P15328) and / or cynomolgus monkey FRα (UniProt ID: A0A2K5U044). In particularly preferred embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof specifically binds to human FRα. In preferred embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof specifically binds to human FRα and cynomolgus monkey FRα.

[0153] In some embodiments, the FRα protein has the sequence of SEQ ID NO: 112 or SEQ ID NO: 113 listed in Figure 8. In one embodiment, the FRα has the sequence of SEQ ID NO: 112 and is a human FRα protein (predicted mature secreted polypeptide). In another embodiment, the FRα has the sequence of SEQ ID NO: 113 and is a cynomolgus monkey FRα protein (predicted mature secreted polypeptide).

[0154] In some embodiments, the ADC antibody or antigen-binding fragment thereof does not bind to one or more selected from mouse FRα (UniProt ID: P35846), rat FRα (UniProt ID: G3V8M6), human FRβ (UniProt ID: P14207), human FRγ (UniProt ID: P41439), or a combination thereof.

[0155] The term "does not bind" means that the ADC antibody or antigen-binding fragment thereof does not substantially bind to one or more of the molecules (e.g., mouse FRα, rat FRα, human FRβ, human FRγ, or a combination thereof). As used herein in the context of binding, the term "substantially free" can mean that less than 5%, 2%, 1%, 0.5%, or 0.1% of cells expressing one or more of the molecules in a cell culture bind (upon contact with) the ADC antibody or antigen-binding fragment thereof. Suitably, as used herein in the context of binding, the term "substantially free" can mean that such cells do not bind.

[0156] D. Antibody preparation ADC antibodies or antigen-binding fragments thereof (e.g., as monoclonal antibodies) can be produced using recombinant DNA methods described in U.S. Patent No. 4,816,567 (herein incorporated by reference). Polynucleotides encoding monoclonal antibodies are isolated from mature B cells or hybridoma cells, such as by RT-PCR using oligonucleotide primers that specifically amplify genes encoding the heavy and light chains of the antibody, and their sequences are determined using conventional procedures. The isolated polynucleotides encoding the heavy and light chains are then cloned into a suitable expression vector, which is transfected into host cells (e.g., E. coli cells, monkey COS cells, Chinese hamster ovary (CHO) cells, or myeloma cells that do not otherwise produce immunoglobulin proteins), and the monoclonal antibodies are produced by the host cells. Alternatively, recombinant monoclonal antibodies or antigen-binding fragments of the desired species can be isolated from phage display libraries expressing the CDRs of the desired species, as described in McCafferty et al., Nature 348:552-554 (1990), Clackson et al., Nature, 352:624-628 (1991), and Marks et al., J. Mol. Biol. 222:581-597 (1991).

[0157] Affinity maturation and chain shuffling strategies are known in the art and can be used to generate high affinity human antibodies or antigen-binding fragments thereof, see Marks et al., BioTechnology 10:779-783 (1992), incorporated by reference in its entirety.

[0158] Various techniques are known for producing antibody fragments. Traditionally, these fragments are derived via proteolytic digestion of intact antibodies, as described, for example, by Morimoto et al., J. Biochem. Biophys. Meth. 24:107-117 (1993) and Brennan et al., Science 229:81 (1985). In some embodiments, ADC antibody fragments are produced recombinantly. Fab, Fv, and scFv antibody fragments can all be expressed and secreted in E. coli or other host cells, thus allowing the production of large amounts of these fragments. Such ADC antibody fragments can also be isolated from the antibody phage libraries described above. ADC antibody fragments can also be linear antibodies, as described in U.S. Pat. No. 5,641,870, incorporated herein by reference. Other techniques for producing antibody fragments will be apparent to those skilled in the art.

[0159] The technique can also be adapted to produce single chain antibodies specific to FRα. See, e.g., U.S. Patent No. 4,946,778. In addition, the method can be adapted to construct Fab expression libraries to allow rapid and efficient identification of monoclonal Fab fragments with the desired specificity for FRα, or derivatives, fragments, analogs, or homologs thereof. See, e.g., Huse et al., Science 246:1275-1281 (1989). Antibody fragments can be produced by techniques known in the art, including, but not limited to, F(ab')2 fragments produced by pepsin digestion of antibody molecules, Fab fragments produced by reducing disulfide bridges of F(ab')2 fragments, Fab fragments produced by treating antibody molecules with papain and a reducing agent, or Fv fragments.

[0160] III. ADC Linkers The ADC antibody or antigen-binding fragment thereof can be conjugated to the ADC payload (e.g., a cytotoxic drug or cytotoxin) by a linker.

[0161] The term "linker" or "spacer" as used herein refers to a bivalent chemical moiety comprising a covalent bond or chain of atoms that covalently attaches an ADC antibody or antigen-binding fragment thereof to an ADC payload (e.g., a cytotoxin) to form an ADC. In some embodiments, the linker or spacer is a peptide spacer. In some embodiments, the linker or spacer is a non-peptide (e.g., chemical) spacer. A suitable linker has two reactive ends, one for antibody conjugation and the other for ADC payload (e.g., a cytotoxin) conjugation. One or both of the reactive ends are absent or incomplete (e.g., only the carbonyl of a carboxylic acid) so that bonds are formed between the linker and / or the ADC payload (e.g., a cytotoxin) and between the linker and / or the ADC antibody or antigen-binding fragment thereof. These conjugation reactions are discussed in more detail below.

[0162] In a preferred embodiment, the linker is cleavably attached (e.g., conjugated) to an amino residue (e.g., an amino acid of an ADC antibody or antigen-binding fragment described herein).

[0163] In some embodiments, the linker is cleavable in the intracellular environment, thereby allowing the Drug Unit to be released from the ADC antibody in the intracellular environment.

[0164] Alternatively, the linker unit may not be cleavable.In such an embodiment, the drug is released, for example, by the degradation of antibody.However, the non-cleavable payload requires complete mAb digestion in lysosome, and the resulting drug-containing product may be too polar, for example, to achieve bystander effect.

[0165] The ADC should preferably be stable and intact before being transported or delivered into the cell, i.e., the antibody is conjugated to the drug moiety. Outside the target cell, the linker is stable, but inside the cell, it can be cleaved at a high rate. An effective linker will: (i) maintain the specific binding properties of the antibody, (ii) allow intracellular delivery of the conjugate or drug moiety, (iii) be stable and intact (i.e., not cleaved) until the conjugate is delivered or transported to its target site, and (iv) maintain the cell-killing or cell-inhibiting effect of the cytotoxic moiety. Standard analytical methods such as mass spectrometry, HPLC, and the separation / analysis technique LC / MS can be used to assess the stability of the ADC.

[0166] The linker can be cleaved, for example, by enzymatic hydrolysis, photolysis, hydrolysis under acidic conditions, hydrolysis under basic conditions, oxidation, disulfide reduction, nucleophilic cleavage, or organometallic cleavage (see, e.g., Leriche et al., Bioorg. Med. Chem., 20:571-582, 2012).

[0167] Linkers that are hydrolyzable under acidic conditions include, for example, hydrazones, semicarbazones, thiosemicarbazones, cis-aconitic amides, orthoesters, acetals, ketals, etc. (See, for example, U.S. Patent Nos. 5,122,368, 5,824,805, and 5,622,929; Dubowchik and Walker, 1999, Pharm. Therapeutics 83:67-123; Neville et al., 1989, Biol. Chem. 264:14653-14661). Linkers that are cleavable under reducing conditions include, for example, disulfides. A variety of disulfide linkers are known in the art, including, for example, disulfide linkers that can be formed using SATA (N-succinimidyl-S-acetylthioacetate), SPDP (N-succinimidyl-3-(2-pyridyldithio)propionate), SPDB (N-succinimidyl-3-(2-pyridyldithio)butyrate), and SMPT (N-succinimidyl-oxycarbonyl-α-methyl-α-(2-pyridyl-dithio)toluene) (see, e.g., Thorpe et al., 1987, Cancer Res. 47:5924-5931; Wawrzynczak et al., In Immunoconjugates: Antibody Conjugates in Radioimaging and Therapy of Cancer (C. W. Vogel ed., Oxford U. Press, 1987)).

[0168] In a preferred embodiment, the linker is susceptible to enzymatic hydrolysis. Such linkers are preferred over pH-sensitive cleavable linkers. pH-sensitive cleavable linkers may not be stable enough and may be prematurely cleaved before reaching the target cell, and thus potential off-target toxicity may be observed. The enzymatically cleavable linker may be a peptide-containing linker that is cleaved by intracellular peptidase or protease enzymes (e.g., including but not limited to lysosomal or endosomal proteases). One advantage of using intracellular proteolytic release of therapeutic agents is that the agent is usually attenuated when conjugated, and the serum stability of the conjugate is high. In some embodiments, the peptidyl linker is at least two amino acids long or at least three amino acids long. Exemplary amino acid linkers include dipeptides, tripeptides, tetrapeptides, or pentapeptides. Peptides containing amino acids including valine, alanine, citrulline (Cit), phenylalanine, lysine, leucine, and glycine are examples of suitable peptides. Natural amino acids, minor amino acids, and non-naturally occurring amino acid analogs (e.g., citrulline) are all examples of amino acid residues that make up the amino acid linker moiety. Exemplary dipeptides include valine-citrulline (VC or Val-Cit) and alanine-phenylalanine (AF or Ala-Phe). Exemplary tripeptides include glycine-valine-citrulline (Gly-Val-Cit) and glycine-glycine-glycine (Gly-Gly-Gly). In some embodiments, the linker comprises a dipeptide such as Val-Cit, Ala-Val, or Phe-Lys, Val-Lys, Ala-Lys, Phe-Cit, Leu-Cit, Ile-Cit, Phe-Arg, or Trp-Cit.

[0169] In some embodiments, the linker comprises PEG. A stable protease cleavable linker containing PEG can limit the hydrophobicity of the payload and selectively cleave and release the free drug in the target cancer cells. The low hydrophobic nature of the linkers described herein allows for high drug loading on the antibody or antigen-binding fragment (e.g., DAR8) without aggregation. This would be significantly higher than mirvetuximab sorafutansine (DAR3-4) or its derivatives, such as IMGN151 (DAR3.5). This allows the ADC to deliver a significantly higher concentration of the ADC payload (e.g., cytotoxin) to the target cancer cells via binding to FRα on the cancer cells.

[0170] In some embodiments, the linker comprises a maleimide. The use of maleimide in the linker can allow the production of DAR8 and DAR4 ADCs by utilizing the natural interchain disulfides in antibodies. This is advantageous over conjugation of surface amines from lysine residues, which can lead to mixed DAR species and batch-to-batch variation. Also, if the conjugation site interferes with antigen binding, there can be reproducibility issues that affect the efficacy of the ADC. Furthermore, other conjugation methods involving engineered antibodies (e.g., azide-alkyne click chemistry) cannot easily achieve a DAR of more than four.

[0171] In certain embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises a linker R L In some embodiments, R is linked to an ADC payload (e.g., a cytotoxin) via L is selected from:

[0172] [C1] [ka]

[0173] In the formula, Q is as follows:

[0174] [C2] [ka]

[0175] In the formula, Q X wherein Q is an amino acid residue, a dipeptide residue, a tripeptide residue, or a tetrapeptide residue, and X in formula 1 is as follows:

[0176] [C3] [ka]

[0177] In the formula, a=0 to 5, b1=0 to 16, b2=0 to 16, c1=0 or 1, c2=0 or 1, d=0 to 5, and at least b1 or b2=0 (i.e., only one of b1 and b2 may not be 0), and at least c1 or c2=0 (i.e., only one of c1 and c2 may not be 0). G in Formula 1 L is a linker for connecting to the ADC antibody or antigen-binding fragment thereof. L is a compound designated "Ib"

[0178] [C4] [ka]

[0179] In the formula, R L1 and R L2 are independently selected from H and methyl, or together with the carbon atom to which they are attached form a cyclopropylene or cyclobutylene group; e is 0 or 1; or a compound designated "Ib'";

[0180] [5] [ka]

[0181] In the formula, R L1 and R L2 is as defined above.

[0182] For example, G L , X, Q X (eg, in the linker of Ia in formula 1) and preferred embodiments of the linker of Ib in formula 4 are outlined.

[0183] The following preferred features may apply to all embodiments or may relate to a single embodiment. The preferred features may be combined together in any combination.

[0184] Various definitions associated with specific terms in this section are provided under the heading "Chemical Definitions" provided below.

[0185] G of the linker of formula 1 L may be selected from the following: [Table 1-1] [Table 1-2]

[0186] In the formula, Ar is C 5-6 represents an arylene group (e.g., phenylene), and X is C 1-4 Represents alkyl.

[0187] In some embodiments, G L is G L1-1 and G L1-2 In some of these embodiments, G L is G L1-1 It is.

[0188] X in formula 1 is preferably as follows:

[0189] [6] [ka]

[0190] In the formula, a=0 to 5, b1=0 to 16, b2=0 to 16, c=0 or 1, d=0 to 5, and at least b1 or b2=0, and at least c1 or c2=0.

[0191] "a" can be 0, 1, 2, 3, 4, or 5. In some embodiments, a is 0 to 3. In some of these embodiments, a is 0 or 1. In further embodiments, a is 0.

[0192] "b1" can be 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 16. In some embodiments, b1 is 0 to 12. In some of these embodiments, b1 is 0 to 8, and can be 0, 2, 3, 4, 5, or 8.

[0193] "b2" can be 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 16. In some embodiments, b2 is 0-12. In some of these embodiments, b2 is 0-8, and can be 0, 2, 3, 4, 5, or 8. Preferably, only one of b1 and b2 may not be 0.

[0194] "c1" can be 0 or 1, and c2 can be 0 or 1. Preferably, only one of c1 and c2 can be non-zero.

[0195] "d" can be 0, 1, 2, 3, 4, or 5. In some embodiments, d is 0-3. In some of these embodiments, d is 1 or 2. In further embodiments, d is 2. In further embodiments, d is 5.

[0196] In some embodiments of X, a is 0, b1 is 0, c1 is 1, c2 is 0, d is 2, and b2 can be 0-8. In some of these embodiments, b2 is 0, 2, 3, 4, 5, or 8. In some embodiments of X, a is 1, b2 is 0, c1 is 0, c2 is 0, d is 0, and b1 can be 0-8. In some of these embodiments, b1 is 0, 2, 3, 4, 5, or 8. In some embodiments of X, a is 0, b1 is 0, c1 is 0, c2 is 0, d is 1, and b2 can be 0-8. In some of these embodiments, b2 is 0, 2, 3, 4, 5, or 8. In some embodiments of X, b1 is 0, b2 is 0, c1 is 0, c2 is 0, and one of a and d is 0. The other of a and d is 1-5. In some of these embodiments, the other of a and d is 1. In other of these embodiments, the other of a and d is 5. In some embodiments of X, a is 1, b2 is 0, c1 is 0, c2 is 1, d is 2, and b1 can be 0-8. In some of these embodiments, b2 is 0, 2, 3, 4, 5, or 8.

[0197] In some embodiments, Q in formula 1 is an amino acid residue. The amino acid can be a natural amino acid or a non-natural amino acid. For example, Q is selected from Phe, Lys, Val, Ala, Cit, Leu, Ile, Arg, and Trp, where Cit is citrulline.

[0198] In some embodiments, Q comprises a dipeptide residue. The amino acids in the dipeptide can be any combination of natural and non-natural amino acids. In some embodiments, the dipeptide comprises natural amino acids. When the linker is a cathepsin-cleavable linker, the dipeptide is the site of action for cathepsin-mediated cleavage. The dipeptide is then the recognition site for the cathepsin.

[0199] In some embodiments, Q is selected from:

[0200] NH -Phe-Lys- C=O ,

[0201] NH -Val-Ala- C=O ,

[0202] NH -Val-Lys- C=O ,

[0203] NH -Ala-Lys- C=O ,

[0204] NH -Val-Cit- C=O ,

[0205] NH -Phe-Cit- C=O ,

[0206] NH -Leu-Cit- C=O ,

[0207] NH -Ile-Cit- C=O ,

[0208] NH -Phe-Arg- C=O ,

[0209] NH -Trp-Cit- C=O , and

[0210] NH -Gly-Val- C=O ,

[0211] wherein Cit is citrulline. Preferably, Q is selected from:

[0212] NH -Phe-Lys-C=O ,

[0213] NH -Val-Ala- C=O ,

[0214] NH -Val-Lys- C=O ,

[0215] NH -Ala-Lys- C=O , and

[0216] NH -Val-Cit- C=O .

[0217] More preferably, Q is NH -Phe-Lys- C=O , NH -Val-Cit- C=O ,or NH -Val-Ala- C=O Other suitable dipeptide combinations include:

[0218] NH -Gly-Gly- C=O ,

[0219] NH -Gly-Val- C=O ,

[0220] NH -Pro-Pro- C=O , and

[0221] NH -Val-Glu- C=O .

[0222] See Dubowchik et al., Bioconjugate Chemistry, 2002, 13, 855-869, incorporated herein by reference.

[0223] In some embodiments, Q is a tripeptide residue. The amino acids in the tripeptide can be any combination of natural and non-natural amino acids. In some embodiments, the tripeptide comprises natural amino acids. If the linker is a cathepsin-cleavable linker, the tripeptide is the site of action for cathepsin-mediated cleavage. The tripeptide is then the recognition site for cathepsin. Particularly interesting tripeptide linkers are the following:

[0224] NH -Glu-Val-Ala- C=O ,

[0225] NH -Glu-Val-Cit- C=O ,

[0226] NH -αGlu-Val-Ala- C=O , and

[0227] NH -αGlu-Val-Cit- C=O .

[0228] In some embodiments, Q is a tetrapeptide residue. The amino acids in the tetrapeptide can be any combination of natural and non-natural amino acids. In some embodiments, the tetrapeptide comprises natural amino acids. When the linker is a cathepsin-cleavable linker, the tetrapeptide is the site of action for cathepsin-mediated cleavage. The tetrapeptide is then the recognition site for the cathepsin. Particularly interesting tetrapeptide linkers are the following:

[0229] NH -Gly-Gly-Phe-Gly C=O , and

[0230] NH -Gly-Phe-Gly-Gly C=O .

[0231] In some embodiments, the tetrapeptide is:

[0232] NH -Gly-Gly-Phe-Gly C=O .

[0233] In the above representation of peptide residues: NH - represents the N-terminus, C=O represents the C-terminus of the residue. The C-terminus is attached to the NH of a "Drug unit" (e.g., A in Formula 14 below). * Glu represents the residue of glutamic acid, i.e.:

[0234] [7] [ka]

[0235] αGlu represents the residue of glutamic acid when bound via the α chain, i.e.:

[0236] [8] [ka]

[0237] In some embodiments, amino acid side chains are chemically protected, if applicable. Side chain protecting groups can be as described above. Protected amino acid sequences can be cleaved by enzymes. For example, dipeptide sequences containing Boc side chain protected Lys residues can be cleaved by cathepsin.

[0238] Protecting groups for the side chains of amino acids are well known in the art and are described in the Novabiochem catalogue, as discussed above.

[0239] R of the linker Ib shown in formula 4 L1 and R L2may be independently selected from H and methyl, or may be taken together with the carbon atom to which they are attached to form a cyclopropylene or cyclobutylene group.

[0240] In some embodiments, R L1 and R L2 are H. In some embodiments, R L1 is H and R L2 is methyl. In some embodiments, R L1 and R L2 Both of the are methyl.

[0241] In some embodiments, R L1 and R L2 together with the carbon atom to which they are attached form a cyclopropylene group. In some embodiments, R L1 and R L2 together with the carbon atom to which they are attached form a cyclobutylene group.

[0242] In group Ib of formula 4, in some embodiments, e is 0. In other embodiments, e is 1 and the nitro group can be at any available position on the ring. In some of these embodiments, it is at the ortho position. In others of these embodiments, it is at the para position.

[0243] In some embodiments, the linker R shown in Formula 1 L is selected from: [Table 2-1] [Table 2-2]

[0244] Preferably, R L is:

[0245] [9] [ka]

[0246] For example, ADCs have the general formula:

[0247] [C10] L-(D L ) p

[0248] or a pharma- ceutically acceptable salt or solvate thereof, where L is an ADC antibody or antigen-binding fragment thereof, and D L is a linker R connected to the ADC antibody or antigen-binding fragment thereof LL In some embodiments, the linker R LL is preferably selected from "Ia'":

[0249] [C11] [ka]

[0250] where Q and X are as defined above, and G LL is a linker attached to the ADC antibody or antigen-binding fragment thereof. Ib' is represented by the formula:

[0251] [C12] [ka]

[0252] In the formula, R L1 and R L2 is as defined above.

[0253] L-(D) of Formula 10 for Antibody Drug Conjugates L ) p"p" in represents the drug loading and is the number of "drug units" (e.g., cytotoxins such as TOPOi) per antibody or antigen-binding fragment thereof. The drug loading ranges from 1 to 20 drug units (D) per antibody or antigen-binding fragment thereof. With respect to the composition, "p" represents the average drug loading of the conjugate in the composition, "p" ranges from 1 to 20. In some embodiments, the range of "p" is selected from 1 to 10, 2 to 10, 2 to 8, 2 to 6, and 4 to 10, preferably p is 8.

[0254] G in Equation 11 LL is selected from: [Table 3-1] [Table 3-2]

[0255] In the formula, Ar is C 5-6 represents an arylene group (e.g., phenylene); X is C 1-4 Represents alkyl.

[0256] In some embodiments, G LL is G LL1-1 and G LL1-2 In some of these embodiments, G LL is G LL1-1 It is.

[0257] In some embodiments, R LL is the above R L It is a group derived from the group.

[0258] It will be recognized by one of skill in the art that any one or more of the chemical groups, moieties, and features disclosed herein can be combined in multiple ways to form linkers useful for conjugating the ADC antibodies and ADC payloads (e.g., cytotoxins) disclosed herein.

[0259] In some embodiments in which the compounds described herein are provided as single enantiomers or enantiomerically enriched forms, the enantiomerically enriched forms have an enantiomeric ratio of greater than 60:40, 70:30, 80:20, or 90:10. In further embodiments, the enantiomeric ratio is greater than 95:5, 97:3, or 99:1.

[0260] IV. Further Chemical Definitions C 5-6 Arylene: As used herein, "C 5-6 The term "arylene" refers to a divalent moiety obtained by removing two hydrogen atoms from an aromatic ring atom of an aromatic compound.

[0261] In this context, prefixes (e.g., C 5-6 ) indicates the number of ring atoms, or range of number of ring atoms, whether carbon atoms or heteroatoms. The ring atoms can be all carbon atoms, as in "carbarylene groups," in which case the group is phenylene (C6). Alternatively, the ring atoms can include one or more heteroatoms, as in "heteroarylene groups." Examples of heteroarylene groups include, but are not limited to, those derived from:

[0262] N1: Pyrrole (azole) (C5), pyridine (azine) (C6),

[0263] O1: Furan (oxol) (C5),

[0264] S1: thiophene (thiol) (C5),

[0265] N1O1: Oxazole (C5), Isoxazole (C5), Isoxazine (C6),

[0266] N2O1: Oxadiazole (Furazan) (C5),

[0267] N3O1: Oxatriazole (C5),

[0268] N1S1: Thiazole (C5), isothiazole (C5),

[0269] N2: imidazole (1,3-diazole) (C5), pyrazole (1,2-diazole) (C5), pyridazine (1,2-diazine) (C6), pyrimidine (1,3-diazine) (C6) (e.g., cytosine, thymine, uracil), pyrazine (1,4-diazine) (C6), and

[0270] N3: Triazole (C5), triazine (C6).

[0271] C 1-4 Alkyl: As used herein, C 1-4 The term "alkyl" refers to a monovalent moiety obtained by removing a hydrogen atom from a carbon atom of a hydrocarbon compound having 1 to 4 carbon atoms, which may be aliphatic or alicyclic, and may be saturated or unsaturated (e.g., partially unsaturated, fully unsaturated). 1-n The term "alkyl" refers to a monovalent moiety obtained by removing a hydrogen atom from a carbon atom of a hydrocarbon compound having 1 to n carbon atoms, which may be aliphatic or alicyclic, saturated or unsaturated (e.g., partially unsaturated, fully unsaturated). Thus, the term "alkyl" includes the subclasses alkenyl, alkynyl, cycloalkyl, etc., discussed below.

[0272] Examples of saturated alkyl groups include, but are not limited to, methyl (C1), ethyl (C2), propyl (C3), and butyl (C4). Examples of saturated straight chain alkyl groups include, but are not limited to, methyl (C1), ethyl (C2), n-propyl (C3), and n-butyl (C4). Examples of saturated branched alkyl groups include isopropyl (C3), isobutyl (C4), sec-butyl (C4), and tert-butyl (C4).

[0273] C 2-4 Alkenyl: As used herein, 2-4The term "alkenyl" refers to an alkyl group having one or more carbon-carbon double bonds. Examples of unsaturated alkenyl groups include, but are not limited to, ethenyl (vinyl, -CH=CH2), 1-propenyl (-CH=CH-CH3), 2-propenyl (allyl, -CH-CH=CH2), isopropenyl (1-methylvinyl, -C(CH3)=CH2), and butenyl (C4).

[0274] C 2-4 Alkynyl: As used herein, "C 2-4 The term "alkynyl" refers to an alkyl group having one or more carbon-carbon triple bonds. Examples of unsaturated alkynyl groups include, but are not limited to, ethynyl (-C≡CH) and 2-propynyl (propargyl, -CH2-C≡CH).

[0275] C 3-4 Cycloalkyl: As used herein, “C 3-4 The term "cycloalkyl" refers to an alkyl group that is also a cyclyl group, i.e., a monovalent moiety obtained by removing a hydrogen atom from an alicyclic ring atom of a cyclic hydrocarbon (carbocyclic) compound, which moiety has 3 to 7 carbon atoms (including 3 to 7 ring atoms). Examples of cycloalkyl groups include, but are not limited to, those derived from saturated monocyclic hydrocarbon compounds, such as cyclopropane (C3) and cyclobutane (C4), and unsaturated monocyclic hydrocarbon compounds, such as cyclopropene (C3) and cyclobutene (C4).

[0276] Connection Label: Where:

[0277] [C13] [ka]

[0278] Superscript Labels C(=O) and NHindicates the group to which the atom is attached. For example, an NH group is shown attached to a carbonyl (not part of the moiety shown) and a carbonyl is shown attached to an NH group (not part of the moiety shown).

[0279] V. ADC cytotoxins The anti-FRα ADCs used in the novel methods disclosed herein comprise an ADC payload and an ADC anti-FRα antibody or antigen-binding fragment thereof. In a preferred embodiment, the anti-FRα ADC comprises an ADC anti-FRα antibody or antigen-binding fragment thereof conjugated to one or more cytotoxins.

[0280] In some embodiments, the number of ADC payloads (e.g., cytotoxins) per antibody (or antigen-binding fragment thereof) can be expressed as the ratio of ADC payloads (i.e., drugs) to antibodies. This ratio is referred to as the drug-to-antibody ratio (DAR). The DAR is the average number of drugs (i.e., ADC payloads) linked to each antibody. In some embodiments, the DAR is in the range of about 1 to 20. In some embodiments, the DAR range is selected from about 1 to 10, about 2 to 10, about 2 to 8, about 2 to 6, and about 4 to 10. In preferred embodiments, the DAR is about 4 (e.g., 3.8 to 4.2) or about 8 (e.g., 7.6 to 8.4), more preferably about 8 (e.g., 7.6 to 8.4).

[0281] Cytotoxins can be any molecule known in the art that inhibits or prevents the function of cells, and / or causes the destruction of cells (cell death), and / or exerts antitumor / antiproliferative effects. Several classes of cytotoxic agents are known to have potential utility in ADC molecules. Suitable cytotoxic agents for use in ADCs include, but are not limited to, topoisomerase I inhibitors (TOPOi), amanitin, auristatin, daunomycin, doxorubicin, duocarmycin, dolastatin, enediyne, lexitropsin, taxane, puromycin, maytansinoid, vinca alkaloid, tubulysin, and pyrrolobenzodiazepines (PBD). Examples of such cytotoxic agents are AFP, MMAF, MMAE, AEB, AEVB, auristatin E, paclitaxel, docetaxel, CC-1065, SN-38, topotecan, morpholino-doxorubicin, rhizoxin, cyanomorpholino-doxorubicin, dolastatin-10, echinomycin, combretastatin, calicheamicin, maytansine, DM-1, vinblastine, methotrexate, and netropsin, and derivatives and analogs thereof. Further disclosure regarding cytotoxins suitable for use in ADCs can be found, for example, in International Patent Application Publication Nos. WO2015 / 155345 and WO2015 / 157592, which are incorporated herein by reference in their entireties.

[0282] In some embodiments, the ADC comprises an anti-FRα antibody or antigen-binding fragment thereof conjugated to one or more cytotoxins selected from a topoisomerase I inhibitor, a tubulysin derivative, a pyrrolobenzodiazepine, or a combination thereof. For example, the ADC antibody or antigen-binding fragment thereof is conjugated to one or more cytotoxins selected from the group consisting of the topoisomerase I inhibitors SG3932 (also known as AZ14170133), SG4010, SG4057, or SG4052 (the structures of which are provided below), or a combination thereof.

[0283] It is preferred that the ADC antibody or antigen-binding fragment thereof is conjugated to a topoisomerase I inhibitor, more preferably the topoisomerase I inhibitor SG3932.

[0284] Topoisomerase inhibitors are compounds that block the action of topoisomerases (topoisomerase I and II), which are a type of enzyme that controls changes in DNA structure by catalyzing the breakage and recombination of the phosphodiester backbone of DNA strands during the normal cell cycle. Topoisomerase I inhibitors are advantageous because they mediate highly effective tumor cell killing with less toxicity to patients. In particular, alternative payloads such as microtubule inhibitors, which are commonly used to date for the development of anti-FRα ADCs, are known to have toxicity issues (Hinrichs, et al. AAPS J. 2015 Sep; 17(5): 1055-1064). Furthermore, the use of less hydrophobic linkers with less potent warheads (e.g., TOPOi) would promote bystander killing in heterogeneous tumors. Bystander activity can be achieved by increasing and / or enhancing the permeability of the warhead by increasing hydrophobicity, but this can result in increased toxicity due to non-specific uptake.

[0285] General examples of suitable topoisomerase I inhibitors are represented by the following compounds:

[0286] [C14] [ka]

[0287] The compound shown above is A * and may be referred to herein as a "drug unit." Compound A *preferably provides a linker for connecting (preferably conjugating) to the ADC antibody or antigen-binding fragment. In a preferred embodiment, the linker is cleavably attached (e.g., conjugated) to an amino residue (e.g., an amino acid of the ADC antibody or antigen-binding fragment).

[0288] More specifically, an example of a suitable topoisomerase I inhibitor is represented by compound "I" having the following formula:

[0289] [C15] [ka]

[0290] Compound I, including salts and solvates thereof, is of the formula: L is a linker for connecting to an ADC antibody or antigen-binding fragment thereof described herein. In some embodiments, R L is as described above.

[0291] Thus, the ADC used in the novel method has the following general formula:

[0292] [C10] L-(D L ) p

[0293] or a pharma- ceutically acceptable salt or solvate thereof, in which L and p are as defined above, L is a topoisomerase I inhibitor having a linker (e.g., drug linker unit) of compound III having the formula:

[0294] [C16] [ka]

[0295] Compound III, including salts and solvates thereof, is of the formula: LLis a linker for connecting to an ADC antibody or antigen-binding fragment thereof described herein. In some embodiments, R LL is as defined above.

[0296] In some embodiments, the compound I of formula 15 has the formula I P is a compound of the type:

[0297] [C17] [ka]

[0298] Compound I P including salts and solvates thereof, wherein R LP is a linker for connecting to an ADC antibody or antigen-binding fragment thereof, the linker being a compound Ia having the formula: P Select from:

[0299] [C18] [ka]

[0300] In the formula, Q P is as follows:

[0301] [C19] [ka]

[0302] In the formula, Q XP Q P is an amino acid residue, a dipeptide residue, or a tripeptide residue. P is as follows:

[0303]

[20] [ka]

[0304] In the formula, aP=0 to 5, bP=0 to 16, cP=0 or 1, and dP=0 to 5. Alternatively, G in formula 18 L is compound Ib having the formula:

[0305]

[21] [ka]

[0306] In the formula, R L1 and R L2 are independently selected from H and methyl, or together with the carbon atom to which they are attached form a cyclopropylene or cyclobutylene group; e is 0 or 1, and aP in formula 20 can be 0, 1, 2, 3, 4, or 5. In some embodiments, aP is 0 to 3. In some of these embodiments, aP is 0 or 1. In further embodiments, aP is 0. bP in formula 20 can be 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 16. In some embodiments, b is 0 to 12. In some of these embodiments, bP is 0 to 8, and can be 0, 2, 4, or 8. cP in formula 20 can be 0 or 1. dP in formula 20 can be 0, 1, 2, 3, 4, or 5. In some embodiments, dP is 0 to 3. In some of these embodiments, dP is 1 or 2. In further embodiments, dP is 2.

[0307] X in Equation 18 P In some embodiments, aP is 0, cP is 1, dP is 2, and bP can be 0 to 8. In some of these embodiments, bP is 0, 4, or 8.

[0308] Q of formula 13 for compound I of formula 15 X A preferred form of the formula is Q XPmay be applied (e.g., where applicable).

[0309] The above G for compound I of formula 15 L , R L1 , R L2 and a preferred form of e is compound I of formula 17 P can be applied to.

[0310] In some embodiments, the conjugate of the ADC of formula 10 is an ADC having the formula: P :

[0311]

[22] L-(D LP ) p

[0312] or a pharma- ceutically acceptable salt or solvate thereof, wherein L is an ADC antibody or an antigen-binding fragment thereof; LP is a compound having the formula III P and R is a topoisomerase I inhibitor (eg, a drug linker unit) designated as:

[0313]

[23] [ka]

[0314] R LLP is a linker connected to the ADC antibody or antigen-binding fragment thereof, the linker being a group Ia P The group Ia' and Ib' are selected from the group Ia' and Ib'. P ' is expressed by the following formula:

[0315]

[24] [ka]

[0316] In the formula, Q P , X P , and G LLis as defined above. The group Ib' is represented by the formula:

[0317]

[25] [ka]

[0318] In the formula, R L1 and R L2 is as defined above, and p is an integer from 1 to 20.

[0319] In some embodiments, the compound I of formula 15 has the formula I P2 Compounds of the type:

[0320]

[26] [ka]

[0321] and salts and solvates thereof. LP2 is a linker for connecting to an ADC antibody or antigen-binding fragment thereof, the linker being a group Ia P2 and Ib. P2 is expressed as follows:

[0322]

[27] [ka]

[0323] In the formula, Q is as follows:

[0324]

[28] [ka]

[0325] In the formula, Q Xis such that Q is an amino acid residue, a dipeptide residue, a tripeptide residue, or a tetrapeptide residue. P2 is as follows:

[0326]

[29] [ka]

[0327] In the formula, aP2=0 to 5, b1P2=0 to 16, b2P2=0 to 16, cP2=0 or 1, dP2=0 to 5, and at least b1P2 or b2P2=0 (i.e., only one of b1 and b2 may not be 0). GL in formula 27 is a linker for connecting to the ADC antibody or antigen-binding fragment thereof.

[0328] The group Ib is represented by the formula:

[0329]

[30] [ka]

[0330] In the formula, R L1 and R L2 are independently selected from H and methyl, or together with the carbon atom to which they are attached form a cyclopropylene or cyclobutylene group; e is 0 or 1.

[0331] X in Equation 27 P2In some embodiments, aP2 can be 0, 1, 2, 3, 4, or 5. In some embodiments, aP2 is 0-3. In some of these embodiments, aP2 is 0 or 1. In further embodiments, aP2 is 0. In some embodiments, b1P2 can be 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 16. In some embodiments, b1P2 is 0-12. In some of these embodiments, b1P2 can be 0-8, 0, 2, 3, 4, 5, or 8. In some embodiments, b2P2 can be 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 16. In some embodiments, b2P2 is 0-12. In some of these embodiments, b2P2 is 0-8, and may be 0, 2, 3, 4, 5, or 8. Preferably, only one of b1P2 and b2P2 may not be 0. In some embodiments, cP2 may be 0 or 1. In some embodiments, dP2 may be 0, 1, 2, 3, 4, or 5. In some embodiments, dP2 is 0-3. In some of these embodiments, dP2 is 1 or 2. In further embodiments, dP2 is 2. In further embodiments, dP2 is 5.

[0332] X P2 In some embodiments, aP2 is 0, b1P2 is 0, cP2 is 1, dP2 is 2, and b2P2 can be 0 to 8. In some of these embodiments, b2P2 is 0, 2, 3, 4, 5, or 8. X P2 In some embodiments, aP2 is 1, b2P2 is 0, cP2 is 0, dP2 is 0, and b1P2 can be 0 to 8. In some of these embodiments, b1P2 is 0, 2, 3, 4, 5, or 8. X P2In some embodiments, aP2 is 0, b1P2 is 0, cP2 is 0, dP2 is 2, and b2P2 can be 0 to 8. In some of these embodiments, b2P2 is 0, 2, 3, 4, 5, or 8. X P2 In some embodiments, b1P2 is 0, b2P2 is 0, cP2 is 0, and one of aP2 and dP2 is 0. The other of aP2 and d is 1-5. In some of these embodiments, the other of aP2 and d is 1. In others of these embodiments, the other of aP2 and dP2 is 5.

[0333] Q of formula 2 for compound I of formula 15 X A preferred embodiment of the invention is the group Ia of formula 27. P2 Q X may be applied (e.g., where applicable) to the above G for compound I of formula 15. L , R L1 , R L2 A preferred embodiment of e is compound I of formula 26 P2 It can be applied to molds.

[0334] In some embodiments, the conjugate of the ADC of formula 10 is a compound ADC having the formula: P2 :

[0335]

[31] L-(D LP2 ) p

[0336] or a pharma- ceutically acceptable salt or solvate thereof, wherein L is an ADC antibody or an antigen-binding fragment thereof; LP2 is a compound having the formula III P2 and a topoisomerase I inhibitor (eg, drug linker unit) which is

[0337]

[32] [ka]

[0338] R LLP2 is a linker connected to the ADC antibody or antigen-binding fragment thereof, the linker being a group Ia P2 The group Ia' and Ib' are selected from the group Ia' and Ib'. P2 ' is expressed by the following formula:

[0339]

[33] [ka]

[0340] In the formula, Q and X P2 is as defined above, and G LL is a linker attached to the ADC antibody or antigen-binding fragment thereof. The group Ib' is represented by the formula:

[0341]

[34] [ka]

[0342] In the formula, R L1 and R L2 is as defined above, and p is an integer from 1 to 20.

[0343] Particularly suitable topoisomerase I inhibitors include those having the formula:

[0344]

[35] [ka]

[0345]

[36] [ka]

[0346]

[37] [ka]

[0347]

[38] [ka]

[0348] and / or

[0349]

[39] [ka]

[0350] The topoisomerase I inhibitor SG3932 is particularly preferred. Thus, in a preferred embodiment, the ADC used in the novel methods comprises an antibody, or antigen-binding fragment thereof, conjugated to a topoisomerase I inhibitor having the formula of SG3932:

[0351]

[35] [ka]

[0352] Synthetic methods for making topoisomerase I inhibitors are described, for example, in WO2020 / 200880, which is incorporated herein by reference.

[0353] As outlined above, topoisomerase I inhibitors are preferred, however, it should be noted that any suitable ADC payload (e.g., drug / cytotoxin) may be linked to the ADC antibody or antigen-binding fragment thereof.

[0354] In certain embodiments, the anti-FRα ADCs used in the invention do not include microtubule inhibitors, such as tubulin inhibitors (e.g., maytansinoids, auristatins). The microtubule inhibitor class of molecules suffers from dose-limiting and potentially difficult-to-treat toxicities.

[0355] VI. Preparation of ADC The ADCs of the disclosure can be made in a variety of ways using known organic chemistry reactions, conditions, and reagents, such as by (1) reacting a reactive substituent of an ADC antibody or antigen-binding fragment with a bivalent linker reagent, which is then reacted with an ADC payload (e.g., a cytotoxin), preferably a topoisomerase I inhibitor, or (2) reacting a reactive substituent of an ADC payload (e.g., a cytotoxin), preferably a topoisomerase I inhibitor, with a bivalent linker reagent, which is then reacted with a reactive substituent of an ADC antibody or antigen-binding fragment thereof.

[0356] As disclosed herein, reactive substituents that may be present in an ADC antibody or antigen-binding fragment thereof include, but are not limited to, (i) an N-terminal amine group, (ii) a side chain amine group, e.g., lysine, (iii) a side chain thiol group, e.g., cysteine, and (iv) a nucleophilic group, such as a sugar hydroxyl group or a sugar amino group, if the antibody is glycosylated. As disclosed herein, reactive substituents that may be present in an ADC antibody or antigen-binding fragment thereof include, but are not limited to, the hydroxyl moieties of serine, threonine, and tyrosine residues, the amino moieties of lysine residues, the carboxyl moieties of aspartic acid and glutamic acid residues, and the thiol moieties of cysteine ​​residues, as well as the propargyl, azido, haloaryl (e.g., fluoroaryl), haloheteroaryl (e.g., fluoroheteroaryl), haloalkyl, and haloheteroalkyl moieties of non-naturally occurring amino acids. In some embodiments, a reactive substituent present in an ADC antibody or antigen-binding fragment thereof disclosed herein comprises an amine moiety or a thiol moiety. Certain antibodies have cysteine ​​bridges that are reducible interchain disulfides. They can be made reactive for conjugation with linker reagents by treating antibodies with reducing agents such as DL-dithiothreitol (DTT) and tris(2-carboxyethyl)phosphine (TCEP). Each cysteine ​​bridge theoretically results in the formation of two reactive thiol nucleophiles. Reaction of lysines with 2-iminothiolane (Traut's reagent), resulting in the conversion of amines to thiols, can be used to introduce additional nucleophilic groups into antibodies. One, two, three, four, or more cysteine ​​residues can be used to insert reactive thiol groups into ADC antibodies (or fragments thereof) (e.g., to prepare mutant antibodies containing one or more non-natural cysteine ​​amino acid residues). Engineered antibodies with reactive cysteine ​​amino acids are described in U.S. Pat. No. 7,521,541 (herein incorporated by reference).

[0357] In some embodiments, the ADC antibody or antigen-binding fragment thereof may have one or more carbohydrate groups that can be chemically modified to contain one or more sulfhydryl groups. The ADC is then formed by conjugation through the sulfur atom of the sulfhydryl group.

[0358] In some embodiments, the ADC antibody may contain one or more carbohydrate groups that can be oxidized to generate aldehyde (-CHO) groups (see, e.g., Laguzza et al., J. Med. Chem. 1989, 32(3), 548-55). Conjugation via the corresponding aldehyde results in the formation of the ADC. Further protocols for the modification of proteins for the attachment or association of ADC payloads (e.g., cytotoxins) are described in Coligan et al., Current Protocols in Protein Science, vol. 2, John Wiley & Sons (2002). Methods for conjugation of linker drug moieties to cell-targeting proteins such as antibodies, immunoglobulins, or fragments thereof can be found, for example, in U.S. Pat. No. 5,208,020, U.S. Pat. No. 6,441,163, WO2005 / 037992, WO2005 / 081711, and WO2006 / 034488, each of which is incorporated herein by reference.

[0359] Conventional conjugation strategies for antibodies or antigen-binding fragments thereof rely on randomly or stochastically conjugating a payload to the antibody or fragment via lysine or cysteine. In some embodiments, the ADC antibody or antigen-binding fragment thereof is stochastically conjugated to an ADC payload (e.g., a cytotoxin), preferably a topoisomerase I inhibitor, for example, by partial reduction of the ADC antibody or fragment followed by reaction with the desired ADC payload with or without the attachment of a linker moiety. The ADC antibody or fragment may be reduced using DTT or other reducing agents to perform a similar reduction (e.g., TCEP). The ADC payload may then be added in molar excess to the reduced antibody or fragment in the presence of DMSO, with or without the attachment of a linker moiety. After conjugation, a quenching agent such as N-acetyl-L-cysteine ​​may be added to quench unreacted drug. The reaction mixture may then be purified (e.g., by TFF, SEC-FPLC, CHT, spin filter centrifugation) and buffer exchanged into PBS or other relevant formulation buffers.

[0360] In some embodiments, the ADC payload (e.g., a cytotoxin) is conjugated to the ADC antibody or antigen-binding fragment thereof by site-specific conjugation. In some embodiments, site-specific conjugation of a therapeutic moiety to the ADC antibody using reactive amino acid residues at specific positions results in a homogenous preparation of the ADC with uniform stoichiometry.

[0361] Site-specific conjugation can be performed via cysteine, residues, or unnatural amino acids. In a preferred embodiment, the ADC payload (e.g., cytotoxin) is conjugated to the ADC antibody or its antigen-binding fragment via at least one cysteine ​​residue. The cysteine ​​amino acid can be engineered at the reactive site in the ADC antibody (or its antigen-binding fragment), and preferably does not form intrachain or intermolecular disulfide bonds (Junutula, et al., 2008b Nature Biotech., 26(8):925-932; Dornan et al. (2009) Blood 114(13):2721-2729; US7521541; US7723485; WO2009 / 052249). In some embodiments, the ADC payload (e.g., a cytotoxin) is conjugated to the antibody or antigen-binding fragment thereof via a cysteine ​​substitution at least one of positions 239, 248, 254, 273, 279, 282, 284, 286, 287, 289, 297, 298, 312, 324, 326, 330, 335, 337, 339, 350, 355, 356, 359, 360, 361, 375, 383, 384, 389, 398, 400, 413, 415, 418, 422, 440, 441, 442, 443, and 446 (numbering corresponds to EU index in Kabat). In some embodiments, the specific Kabat positions are 239, 442, or both. In some embodiments, the specific position is Kabat position 442, an amino acid insertion between Kabat positions 239 and 240, or both. In some embodiments, the ADC payload (e.g., a cytotoxin) is conjugated to the ADC antibody or antigen-binding fragment thereof via a thiol-maleimide bond. In some aspects, the amino acid side chain is a sulfhydryl side chain.

[0362] When multiple nucleophilic or electrophilic groups of an ADC antibody or antigen-binding fragment thereof react with an ADC payload (e.g., a cytotoxin), the resulting product may be a mixture of ADCs with a distribution of ADC payload units (e.g., one, two, three, etc.) attached to the antibody. Liquid chromatography methods such as hydrophobic interaction (HIC) can separate compounds in a mixture by ADC payload loading value. Preparations of ADCs with a single ADC payload loading value (p) can be isolated.

[0363] The average number of ADC payloads (e.g., cytotoxins) per ADC antibody (or antigen-binding fragment) in a preparation of ADC from a conjugation reaction can be characterized by conventional means such as UV, reversed-phase HPLC, HIC, mass spectrometry, ELISA assay, and electrophoresis. The quantitative distribution of ADCs in units of p can also be determined. By ELISA, the average value of p in a particular preparation of ADC can be determined (Hamblett et al. (2004) Clin. Cancer Res. 10:7063-7070, Sanderson et al. (2005) Clin. Cancer Res. 11:843-852). In some cases, separation, purification, and characterization of homogeneous ADCs with a particular value of p from antibodies with other ADC payloads (e.g., cytotoxins) can be achieved by means such as reversed-phase HPLC, electrophoresis, TFF, SEC-FPLC, CHT, spin filter centrifugation, etc. Such techniques are also applicable to other types of conjugates.

[0364] VII. Pharmaceutical Compositions The anti-FRα ADC can be administered as a pharmaceutical composition containing one or more pharma- ceutically compatible components.

[0365] The term "pharmaceutical composition" refers to a preparation that is in a form that allows the biological activity of the active ingredient to be effective and does not contain additional ingredients that are unacceptably toxic to the subject to which the composition is administered. Such compositions may be sterile and may contain a pharma-ceutically acceptable carrier, such as physiological saline. Suitable pharmaceutical compositions may include one or more of buffers (e.g., acetate, phosphate, or citrate buffers) to enhance bioavailability, surfactants (e.g., polysorbates), stabilizers (e.g., human albumin), preservatives (e.g., benzyl alcohol), and absorption enhancers, and / or other conventional solubilizers or dispersants.

[0366] As used herein, the term "pharmaceutical acceptable" means approved by a regulatory agency of the Federal or state government, or listed in the United States Pharmacopoeia, the European Pharmacopoeia, or other generally recognized pharmacopoeias for use in animals, especially humans.

[0367] In some embodiments, the pharmaceutical composition may be contained within one or more formulations selected from a capsule, a tablet, an aqueous suspension, a solution, a nasal aerosol, a lyophilized powder that can be reconstituted to create a suspension or solution prior to use, or combinations thereof.

[0368] For example, the anti-FRα ADC AB1370049-SG3932-DAR8 used in the novel methods can be administered as a pharmaceutical composition containing a buffer, such as a histidine buffer, an excipient, such as sucrose, and a surfactant, such as polysorbate 80. Pharmaceutical compositions containing the anti-FRα ADC used in the novel methods can be used as injections, either as aqueous injections or lyophilized injections, or even as lyophilized injections.

[0369] The pharmaceutical compositions disclosed herein can be administered to a patient by any suitable systemic or local route of administration.For example, administration can be oral, buccal, sublingual, ophthalmic, intranasal, intratracheal, pulmonary, topical, transdermal, genitourinary, rectal, subcutaneous, intravenous, intraarterial, intraperitoneal, intramuscular, intracranial, intrathecal, epidural, intraventricular, or intratumoral.

[0370] The pharmaceutical compositions disclosed herein can be formulated for administration by any suitable means, for example, by epidermal or transdermal patch, ointment, lotion, cream, or gel, by sprayer, vaporizer, or inhaler, by injection or infusion, or in the form of capsules, tablets, solutions or suspensions in water or non-aqueous media, drops, suppositories, enemas, sprays, or powders. The most suitable route for administration in any given case will depend on the physical and mental condition of the patient, the nature and severity of the disease, and the desired properties of the formulation.

[0371] The pharmaceutical composition for oral administration may be in the form of tablet, capsule, powder or liquid. Tablets may contain solid carriers or adjuvants. Liquid pharmaceutical compositions generally contain liquid carriers such as water, petroleum, animal or vegetable oils, mineral oil or synthetic oil. They may contain saline, dextrose or other sugar solutions, or glycols such as ethylene glycol, propylene glycol or polyethylene glycol. Tablets may contain solid carriers such as gelatin.

[0372] For intravenous, cutaneous or subcutaneous injection, or injection at the affected site, the active ingredient is in the form of a parenterally acceptable aqueous solution, which is pyrogen-free and has suitable pH, isotonicity and stability.Those skilled in the art can prepare suitable solutions using isotonic vehicles such as, for example, sodium chloride injection, Ringer's injection, lactated Ringer's injection, etc.Preservatives, stabilizers, buffers, antioxidants, and / or other additives may be included as necessary.

[0373] When the pharmaceutical composition containing the anti-FRα ADC used in the novel methods is an aqueous injection solution, it can be diluted with a suitable diluent and then administered as an intravenous infusion. Examples of diluents can include dextrose solution and saline.

[0374] When the pharmaceutical composition containing the anti-FRα ADC used in the novel method is a lyophilized injectable, it is dissolved in injection-grade water, then diluted with a suitable diluent to the required amount, and then administered as an intravenous infusion. Examples of diluents include dextrose solution and saline.

[0375] VIII. Therapeutic Uses of Anti-FRα ADCs Preferably, the ADCs disclosed herein can be used to prevent, treat, or ameliorate symptoms associated with a disease, disorder, or infection, preferably, cancer. The term "cancer" is used interchangeably with the term "tumor."

[0376] "Treating" refers to a therapeutic measure that cures, slows, alleviates symptoms, and / or stops the progression of a diagnosed pathological condition or disorder. Thus, a person in need of treatment includes a person who already has the disorder. In some embodiments, a subject is successfully "treated" for a disease or disorder (preferably cancer) according to the novel methods disclosed herein when the patient exhibits, for example, complete, partial, or transient alleviation or elimination of symptoms associated with the disease or disorder (preferably cancer).

[0377] "Preventing" refers to prophylactic or preventative measures that prevent and / or delay the onset of a targeted pathological condition or disorder, preferably cancer. Thus, those in need of prevention include those prone to having the disorder or susceptible to the disorder.

[0378] The terms "patient", "individual" and "subject" are used interchangeably herein to refer to a mammalian subject. In some embodiments, the "subject" is a human, a farm animal, a livestock animal, a sport animal, and a zoo animal, such as a human, a non-human primate, a dog, a cat, a guinea pig, a rabbit, a rat, a mouse, a horse, a cow, and the like. In some embodiments, the subject is a cynomolgus monkey (Macaca fascicularis). In a preferred embodiment, the subject is a human. In some embodiments of the novel method, the patient may not have been previously diagnosed with cancer. Alternatively, the patient may have been previously diagnosed with cancer. The patient may also be a patient who exhibits disease risk factors or is asymptomatic for cancer. The patient may also be a patient who is suffering from cancer or at risk for developing cancer. Thus, in one embodiment, the method of the present invention can be used to confirm the presence of cancer in a patient. For example, the patient may have been previously diagnosed with cancer by alternative means. In one embodiment, the patient has previously been administered cancer therapy.

[0379] In some embodiments, the anti-FRα ADCs disclosed herein can be used to treat cancers associated with FRα expression. In other words, the cancers referred to herein can include cancerous cells that express FRα. The cancer cells can be contained within a tumor. In some embodiments, the cancer includes cancer cells that have heterogeneous expression of FRα and / or low expression of FRα.

[0380] After undergoing cellular internalization in cancer cells, the anti-FRα ADC used in the present invention is cleaved by the linker unit, releasing the drug unit into the cancer cells. The anti-FRα antibody-drug conjugate used in the novel method also has a bystander effect, in that the anti-FRα antibody-drug conjugate is internalized into cancer cells expressing the target protein FRα, and then the drug unit exerts an antitumor effect on adjacent cancer cells that do not express the target protein FRα.

[0381] Preferably, the cancer is selected from ovarian cancer, lung cancer, endometrial cancer, breast cancer (e.g., TNBC), cervical cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), and malignant pleural mesothelioma. More preferably, the cancer is ovarian cancer or lung cancer. In some embodiments, the cancer is one or more of non-small cell lung cancer (NSCLC), preferably selected from squamous cell NSCLC, adenocarcinoma NSCLC, or a combination thereof.

[0382] Further examples of cancer include, but are not limited to, benign, pre-malignant, and malignant cell proliferations, where malignant cell proliferations include, but are not limited to, neoplasms and tumors (e.g., histocytoma, glioma, astrocytoma, osteoma), cancer (e.g., ovarian cancer, lung cancer, non-small cell lung cancer (squamous cell carcinoma or adenocarcinoma), endometrial cancer, pancreatic cancer, gastric cancer, colon cancer, squamous cell carcinoma of the head and neck, malignant pleural mesothelioma, breast cancer (e.g., TNBC), and renal cancer. Any type of cell can be treated, including, but not limited to, lung, gastrointestinal, breast (mammary), ovarian, kidney (renal), and pancreatic.

[0383] The anti-FRα ADCs disclosed herein are expected to exert a therapeutic effect when applied to patients as systemic therapy, and further when applied locally to cancer tissues.

[0384] IX. Diagnostic Antibodies The diagnostic antibody used in the novel method binds to FRα on cells (e.g., cancer cells) in tissue samples from cancer patients. In preferred embodiments, the diagnostic antibody specifically binds to human FRα. The diagnostic antibody is labeled to aid in detection of cell binding. In some embodiments, the label can be a fluorophore or dye. In some embodiments, the label can be used for immunohistochemistry. In preferred embodiments, the diagnostic antibody is linked to 3,3′-diaminobenzidine (DAB). In some embodiments, the diagnostic antibody is an anti-FRα antibody (VMSI, #742-5065) obtained from Ventana. In some embodiments, the diagnostic antibody is an anti-rabbit FRα antibody. In preferred embodiments, the diagnostic antibody is an anti-rabbit FRα antibody (#ab221543) obtained from Abcam. In some embodiments, the diagnostic antibody is closely related to the ADC antibodies or antigen-binding fragments described herein. In some embodiments, the diagnostic antibody is closely related to the ADC antibody of ADC AB1370049-SG3932.

[0385] In some embodiments, the diagnostic antibody binds to cancer cells from ovarian cancer, lung cancer, endometrial cancer, colon cancer, breast cancer (e.g., TNBC), cervical cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), and malignant pleural mesothelioma. More preferably, the diagnostic antibody binds to cancer cells from ovarian cancer, lung cancer, endometrial cancer, or breast cancer. In some embodiments, the diagnostic antibody binds to cancer cells from non-small cell lung cancer (NSCLC), preferably selected from squamous cell NSCLC, adenocarcinoma NSCLC, or a combination thereof.

[0386] X. Methods for predicting patient response to administration of anti-FRα ADCs FIG. 9 is a flow chart of steps 11-17 of method 10, in which an analysis system analyzes digital images of tissue from a cancer patient to predict how the cancer patient will respond to a therapy involving an anti-FRα antibody drug conjugate (ADC). In one embodiment, the ADC is anti-FRα ADC AB1370049-SG3932-DAR8. In one embodiment, the method predicts the response of a patient with a cancer selected from ovarian cancer, lung cancer, endometrial cancer, breast cancer (e.g., TNBC), cervical cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), and malignant pleural mesothelioma to an ADC. More preferably, the method predicts the response of a patient with a cancer selected from ovarian cancer, non-small cell lung cancer, endometrial cancer, and colon cancer to an ADC. In some embodiments, the methods predict response to an ADC in a patient having non-small cell lung cancer (NSCLC), preferably selected from squamous NSCLC, adenocarcinoma NSCLC, or a combination thereof.

[0387] The first step 11 involves obtaining high resolution digital images of tissue slices from a cancer patient that have been stained using one or more biomarkers or stains.

[0388] To predict the efficacy of an ADC therapy, a diagnostic antibody (e.g., a diagnostic biomarker) with an attached dye that targets the same protein as the protein targeted by the ADC therapy is used. In one embodiment, the anti-FRα ADC therapy to be scored is an anti-FRα antibody conjugated to a topoisomerase I inhibitor.

[0389] Thus, in embodiments, the diagnostic antibody also targets the FRα protein.

[0390] In step 12, the pre-trained convolutional neural network processes digital images of cancer patient tissue stained with diagnostic antibodies linked to dyes such as 3,3'-diaminobenzidine (DAB). The staining intensity of the dye in the membrane of the cancer cell is determined based on the average staining intensity of the dye of all pixels associated with the corresponding segmented membrane object. Furthermore, the staining intensity of the dye at a single pixel is calculated based on the pixel's red, green, and blue color components. The result of the image analysis process is two posterior image layers that represent, for each pixel in the digital image, the probability that the pixel belongs to a cell nucleus and the probability that the pixel belongs to a cell membrane.

[0391] In step 13, individual cancer cells are detected based on posterior layer heuristic image analysis for nuclei and membranes. Cancer cell objects including cell membrane objects are generated.

[0392] In step 14, each cancer cell is identified as either optical density positive or optical density negative based on the amount of DAB in the cell membrane. The amount of DAB is determined by the staining intensity of each membrane based on the mean (average) optical density of the brown diaminobenzidine (DAB) signal in all pixels of the membrane. Each cancer cell is identified as either (i) optical density positive if the mean optical density of the cancer cells is equal to or greater than the optical density threshold, or (ii) optical density negative if the mean optical density of the cancer cells is less than the optical density threshold. In one embodiment, the optical density threshold is in the range of 10-15 on a scale of 220 maximum optical density. For example, the optical density threshold is 12.

[0393] In step 15, a binary proximity score is generated for the digital image of the tissue sample that is equal to the percentage of cancer cells in the digital image that are either optical density positive or optical density negative, but within a predetermined distance from optical density positive cancer cells. In one embodiment, the predetermined distance is 25 microns.

[0394] In step 16, a cancer patient is identified as one who would likely benefit from administration of an anti-FRα ADC if the proximity score exceeds a predetermined percentage threshold. In one embodiment, the predetermined percentage threshold is 20.

[0395] The threshold optical density used in step 14, the predefined distance used in step 15, and the predefined percentage threshold used in step 16 are optimized using a training cohort of patients with known response to ADC therapy. Optimization is performed using Spearman correlation analysis of the patient's proximity score versus response to obtain the lowest p-value and Rank R-value. Alternatively, a receiver operating characteristic (ROC) analysis is performed for cut-point optimization to distinguish binary clinical outcomes (e.g., responsive R / non-responsive NR). The binary proximity score indicates how a cancer patient will respond to therapy with an anti-FRα ADC.

[0396] In step 17, if the score exceeds a predetermined percentage threshold, therapy with an anti-FRα ADC is recommended for patients who score positive.

[0397] XI. Examples of Prediction and Scoring Methods A. Image analysis of stained tissue The method of FIG. 9 will now be described with reference to a particular image of stained cancer tissue.

[0398] In step 11, the tissue sample is immunohistochemically stained using a dye linked to a diagnostic antibody that binds to an associated protein on cancer cells in the tissue sample. FIG. 10 (top left image) is a digital image 18 of a portion of the stained tissue obtained in step 11. Image 18 shows tissue from a cancer patient immunohistochemically stained with an anti-FRα diagnostic antibody linked to a dye. In this example, the diagnostic antibody is an anti-FRα generating clone (designated FRα IHC Tool Ab) reformatted as a mouse anti-human FRα IgG1 clone. IgG1 indicates the isotype of the anti-FRα antibody. The staining protocol was developed using the Dako Autostainer platform. The anti-FRα antibody binds to the membrane protein FRα. Therefore, 3,3′-diaminobenzidine (DAB) staining indicates the location of the protein FRα in the tissue sample.

[0399] In step 12, image analysis is performed on the digital image 18 to generate a posterior image layer of the nuclei and membranes of the cancer cells using a convolutional neural network. Image analysis is used to detect the cancer cells and their components (e.g., nuclei, membranes, and cytoplasm). Figure 10 illustrates the image analysis process of step 12. The convolutional neural network generates a posterior layer (grey value image) that indicates for each pixel of the digital image 18 the probability that each pixel belongs to either the nucleus (top right image of Figure 10) or the membrane (bottom left image of Figure 10) of the cell. High probabilities are shown in black and low probabilities in white. The bottom left image also shows the nuclear center being identified first to generate the cell membrane object by extending outward from the nuclear center until it meets an extension from the center of another nuclear.

[0400] In one embodiment, the convolutional neural network includes a series of convolutional layers from the input image 18 to a bottleneck layer with a very low spatial size (1-16 pixels) and a series of deconvolutional layers to backward layers with the same size as the input image 18. This network architecture is called U-Net. The weights of the convolutional neural network are trained by generating manually annotated layers of nuclei and membranes in multiple training images, and then adjusting the network weights by an optimization algorithm so that the generated backward layers are most similar to the manually generated annotation layers.

[0401] In another embodiment, the annotation layers of nuclei and membranes are automatically generated and manually corrected in multiple training images. Epithelial regions and nuclei centers are manually annotated as regions and points, respectively. For each training image, a membrane segmentation is automatically generated by applying a region-growing-like algorithm (e.g., watershed segmentation) seeded by the annotated nuclei centers and constrained by the extent of the annotated epithelial regions. Given a training image, a nuclei segmentation is automatically generated by applying a blob detection algorithm (e.g., by the Maximum Stable Extremum Region MSER algorithm) and selecting as nuclei only those detected blobs that contain annotated nuclei centers. The automatically generated membrane and nuclei segmentations are visually confirmed and manually corrected if necessary. The correction step involves one of rejecting incorrectly segmented membranes or nuclei, explicitly accepting correctly annotated membranes or nuclei, or refining the shape of the membrane or nucleus. For each image with annotated membranes or nuclei, an annotation layer is created. In one embodiment, each pixel in the annotation layer is assigned a "1" if it belongs to an annotated object (membrane or nucleus) and a "0" if it does not. In another embodiment, the pixels in the annotation layer represent the distance to the nearest annotated object. The network weights are adjusted by an optimization algorithm so that the generated posterior layers are most similar to the automatically generated membrane and nucleus annotation layers.

[0402] FIG. 11 illustrates step 13, where individual cancer cell objects, each containing a cell membrane and cytoplasm, are detected. The heuristic image analysis process uses watershed segmentation to segment cell nuclei using a nuclear posterior layer generated by a convolutional neural network. The segmentation produces nuclear objects. Each nuclear object is assigned a unique identifier (UID). Individually identified nuclei are shown as dark objects in FIG. 11 (bottom right image). Detected nuclei are also displayed as overlays on the input image 18 (top left image) and the posterior layer of the nuclei (top right image) and membrane (bottom left image).

[0403] In one embodiment, watershed segmentation involves thresholding the nuclear posterior layer with a predefined first size threshold. All single connected pixels above the first size threshold are considered to belong to a nuclear object. Nuclear objects with an area less than 16um^2 are discarded. A UID is assigned to each nuclear object. In a subsequent step, the nuclear object is expanded towards the smaller nuclear posterior, and the added nuclear posterior pixels must be larger than a second predefined threshold.

[0404] FIG. 12 shows the detection and segmentation of membrane objects, where membrane objects are segmented by expanding the region of boundary pixels of detected cells outward to the membrane probability layer and a predefined membrane layer posterior threshold. Thicker boundary regions become membrane objects. Each membrane object is assigned the same UID as that of its associated nucleus object.

[0405] The space between the membrane and the nucleus is assigned to the cytoplasm using the UID of the nucleus. For each membrane (see image at top left in Figure 12) and cytoplasm (see image at top left in Figure 12), the average optical density of the DAB staining is exported to a file on the hard drive along with the UID. For each cell (defined to include nucleus, cytoplasm, and membrane), the location of the cell's centroid (x,y) within the slide is also exported. The files may reside on a hard disk, a solid state disk, or a portion of dedicated RAM within the computer system.

[0406] FIG. 13 shows the results of image analysis in an image analysis software environment. FIG. 13 (top left image) shows the segmentation of nuclear and membrane objects as an overlay on a digital image of stained tissue. FIG. 13 (bottom left image) shows the segmentation of nuclear objects as an overlay on an optical density representation of the digital image. Dark optical density pixels are associated with a large amount of DAB and light optical density pixels are associated with a small amount of DAB. The DAB optical density of each image pixel is calculated from the red-green-blue representation of the image pixel by transforming the red-green-blue color space such that the brown DAB component is an independent color and by taking the logarithm of that brown color component. FIG. 13 (top right image) shows the image analysis script used to generate the segmented image. FIG. 13 (bottom right image) shows the exported measurements of all cell membrane and cytoplasmic objects in image 18.

[0407] In another embodiment, the cell membrane of each tumor cell is detected by the single-shot instance segmentation method "StarDist" as described in [https: / / squidpy.readthedocs.io / en / stable / external_tutorials / tutorial_stardist.html]. This method approximates the cell membrane as a closed polygon with N star-like support vectors originating from the cell center. The network predicts N+1 posterior layers, where layer N+1 associates each pixel value with the probability that this pixel belongs to the cell center. In each of the N posterior layers, the distance of the membrane to the cell center pixel is predicted. The average optical density of each cell membrane is calculated by iterating through the predicted closed polygon and measuring the optical density of the associated pixels that the polygon crosses.

[0408] B. Calculating Predictive Validity Scores Based on the optical density of the DAB staining in the membrane objects and optionally cytoplasmic objects, a spatial proximity score is determined for the tissue sample shown in the digital image 18. The spatial proximity score also accounts for the staining intensity of the DAB dye in the membrane objects and optionally cytoplasmic objects of adjacent cancer cells that are closer than a predetermined distance to the cancer cell for which the single cell ADC score is calculated. There are two types of spatial proximity scores: binary spatial proximity score (bSPS) and continuous spatial proximity score (cSPS).

[0409] 14 shows an example of how the binary spatial proximity score (bSPS) is determined. The binary spatial proximity score is a percentage score equal to the sum of (i) the number of cancer cells in the digital image 18 that have a staining intensity equal to or greater than the optical density threshold, and (ii) the number of cancer cells in the digital image that have a staining intensity less than the optical density threshold but within a predetermined distance of at least one cancer cell that has a staining intensity equal to or greater than the optical density threshold, divided by the total number of cancer cells in the digital image 18.

[0410] FIG. 14 shows ten exemplary cancer cells (also called tumor cells) with various optical densities of staining from the image analysis of steps 12-13. The steps of heuristic image analysis shown in FIGS. 10-12 are used to obtain a segmentation into image objects including cell membrane objects and cancer cell objects. Circles with solid perimeters represent tumor cells whose optical density of the brown DAB staining (corresponding to the amount of target protein FRα) is equal to or greater than a predefined optical density threshold. In this example, the optical density threshold is 12 out of a maximum scale of optical density of 220. Three cancer cells 19-21 in FIG. 14 are classified as optical density positive. Circles with dashed perimeters represent tumor cells whose optical density of the brown DAB staining is less than a predefined optical density threshold. Seven cancer cells 22-28 are classified as optical density negative. Darker grey circles with dashed perimeters represent optical density negative cancer cells located within a predefined distance from at least one optical density positive cancer cell. Two cancer cells 22-23 are optical density negative and are within a predetermined distance from an optical density positive cancer cell, in this example the predetermined distance is 25 microns.

[0411] In this example, the binary spatial proximity score is equal to a percentage score of 50%, which is calculated as the sum of three optical density positive cells and two optical density negative cells within a given distance of an optical density positive cell divided by 10 cancer cells ([3+2] / 10=0.5), as shown in FIG. 14.

[0412] In the schematic image of Figure 14, three cancer cells 19-21 express high amounts of the target protein FRα and would be very likely to be killed by the ADC payload (e.g., cytotoxin) entering the cell linked to the ADC antibody (effect 1 in Figure 15). Two cancer cells 22-23 do not express sufficient amounts of the target protein FRα to be directly killed by the anti-FRα ADC. However, due to the proximity of at least one of the optical density positive cells 19-21, the toxic payload released from cells 19-21 would also kill cancer cells 22-23 (effect 3 in Figure 15). The remaining optical density negative cancer cells 24-28 remain active and may be the origin of drug resistance mechanisms, which may ultimately cause the death of the patient.

[0413] Figure 15 shows the mechanism by which anti-FRα ADC therapy kills cancer cells. In the first step, the ADC antibody binds to the target protein FRα and inhibits the natural function of the target protein, which can lead to cell death. In the second step, the payload (e.g., a type I topoisomerase inhibitor) is internalized into the cell and kills the cell due to the toxicity of the payload. This uptake of the payload depends on the amount of the target protein on the membrane and also on the difference in the amount of the target protein on the membrane and in the cytoplasm. After being uptaken, the payload can be released from the cell into the surrounding tissue. In the third step, the payload can enter nearby cells and kill them as well. The spatial distribution of the payload in the tissue spreads by passive diffusion.

[0414] Conventional IHC scoring reflects both the effect of target protein inhibition by ADC binding as well as the effect of cytotoxic payload entering cancer cells with the ADC antibody. Therefore, conventional scoring of ADC therapy does not reflect the importance of the presence of target protein in the cytoplasm and the effect of cytotoxic payload diffusing into tissues after being released from the initially killed cancer cell. In comparison, the novel predictive spatial proximity score measures the effect of cytotoxic payload release on neighboring cancer cells.

[0415] 16-18 show an example of how the continuous spatial proximity score is determined. Although the continuous spatial proximity score applies a predetermined distance, optical density negative cancer cells still receive some weight even though they are farther than a predetermined distance from optical density positive cancer cells. In addition, this embodiment of the continuous spatial proximity score also takes into account staining in the cytoplasm. The continuous spatial proximity score is calculated for the tissue sample shown in digital image 18 based on the optical density of DAB staining within membrane objects and optionally cytoplasmic objects.

[0416] FIG. 16 shows an exemplary quantitative result of the optical density of the staining from the image analysis of steps 12-13 in the schematic diagram using the grey values ​​of the membrane and cytoplasmic pixels. The heuristic image analysis steps shown in FIGS. 10-12 are used to obtain the exemplary segmentation of FIG. 16 into the cell nucleus, cell membrane, and cytoplasm. The light grey values ​​in FIG. 16 are associated with high DAB optical density and therefore high amounts of the protein targeted by the diagnostic antibody. The dark grey values ​​are associated with low DAB optical density. Lighter pixels correspond to higher DAB optical density.

[0417] FIG. 17 lists exemplary quantitative amounts of staining on the membrane and in the cytoplasm of the images of FIG. 16, some of which are reproduced in FIG. 17. The optical densities of the brown DAB signal from the membrane of the first, second, and third cells are 0.949, 0.369, and 0.498, respectively. In this example, the optical densities are expressed as a percentage of the maximum staining of 255. The optical densities of the brown DAB signal from the cytoplasm of the first, second, and third cells are 0.796, 0.533, and 0.369, respectively. In the schematic image of FIG. 17, the first cancer cell 29 expresses a large amount of the target protein FRα and would be very likely to be killed by the ADC payload (e.g., cytotoxin) that enters the cell linked to the ADC antibody (effect 1 in FIG. 15). The second cancer cell 30 and the third cancer cell 31 do not express sufficient amounts of the target protein FRα to be directly killed by the anti-FRα ADC. However, due to the proximity of the second cancer cell 30 to the first cancer cell 29, the toxic payload released from the first cancer cell will also kill the second cancer cell 30 (effect 3 in FIG. 15 ). The third cancer cell 31 will remain active and may be the origin of drug resistance mechanisms, which may ultimately cause the death of the patient.

[0418] In step 14 of the method of FIG. 9, the average optical density of each cancer cell is determined. Then, in a modification of method 10, a single-cell spatial proximity score is determined. FIG. 18 shows the calculation of the single-cell spatial proximity score for each of the three cells shown in FIG. 17, incorporating an exponential weighting factor of cell separation to reflect the uptake of the ADC payload (e.g., cytotoxin) into adjacent cells. The single-cell score can be calculated based on the formula shown in FIG. 18. The predetermined distance used in the formula is 50 microns. The optical densities listed in FIG. 17 for DAB signals from the cell membrane (0.949, 0.369, and 0.498) and cytoplasm (0.796, 0.533, and 0.369) are input into the calculation shown in FIG. 18. The first cell 29, the second cell 30, and the third cell 31 have single-cell spatial proximity scores of 0.145, 0.012, and 0.064, respectively.

[0419] Thus, the continuous spatial proximity score incorporates a measure of the amount of target protein on the cell membrane using DAB optical density, and optionally an estimate of ADC payload (e.g., cytotoxin) uptake. As shown in FIG. 17, the uptake of the ADC payload (e.g., cytotoxin) of a first cell depends on both the amount of dye in its membrane and its cytoplasm, and the amount of dye in the membrane and cytoplasm of a second cell nearby the first cell. More specifically, the neighborhood can be a disk with a predetermined radius around the first cancer cell. In one embodiment, the continuous spatial proximity score of a first cancer cell is determined by a distance-weighted sum of the powers of the DAB optical density of membrane and cytoplasmic objects whose associated cancer cell center is closer to the first cancer cell center than a predetermined distance. In one embodiment, the predetermined distance is 50 μm, as used in the calculation of FIG. 18. In another embodiment, the distance is 20 μm. In another embodiment, the distance is determined by minimizing the Kaplan Meier log-rank p-value of the spatial proximity score positive arm compared to the spatial proximity score negative arm in a given training cohort of ADC-treated patients. In yet another embodiment, distance weighting involves calculating the exponential of the sum of the scaled negative Euclidean distances from the first cancer cell center to the other cancer cell centers, hi another embodiment, the powers of the sum are restricted to 0, 1, and 2 (constant, linear, squared).

[0420] C. Validation of binary spatial proximity scores The accuracy of the predicted ADC scores generated according to the novel method and its variants in Figure 9 was validated based on patient-derived xenograft (PDX) studies of patients with various FRα expression levels and correlated with tumor growth as an indicator of response to ADC. Tumor tissue fragments were implanted subcutaneously into 6-8 week old female athymic nude mice. When tumors reached an appropriate tumor volume range (typically 150-300 mm), animals were randomized into treatment and control groups and administration of anti-FRα ADC AB1370049-SG3932-DAR8 was initiated. Doses of 2.5 mg / kg and 5 mg / kg were used. The human equivalent doses (HED) are 0.2 mg / kg and 0.4 mg.kg, respectively. Tumor-bearing mice were administered a single dose of anti-FRα ADC via intravenous injection. Mice were observed daily and tumor dimensions and body weights were measured and recorded twice weekly. Tumor volumes were measured by digital calipers and tumor volumes were calculated using the following formula: Tumor volume = [length (mm) x width (mm) x 0.52 (where length and width are the long and short diameters of the tumor, respectively).

[0421] In step 15 of method 10, a predictive validity score in the form of a binary spatial proximity score (bystander_memb(meanOD)_binary_r25_cut12) is generated for the tissue sample based on the percentage of cancer cells in the digital image that are either optical density (OD) positive or optical density negative, but within a predetermined distance from an optical density positive cancer cell. In this embodiment, the predetermined distance is 25 microns. Each cancer cell is identified as optical density positive if its mean optical density is equal to or greater than an optical density threshold, and as optical density negative if its mean optical density is less than the optical density threshold. In this embodiment, the optical density threshold is 12 on a scale of 220 maximum optical density.

[0422] In step 15 of method 10, a cancer patient is identified as one who would likely benefit from administration of an ADC if the binary spatial proximity score exceeds a predetermined threshold. In this embodiment, the predetermined threshold is 98 percent. The predetermined distance, optical density threshold, and predetermined threshold correlate with the response of training patients treated with the ADC.

[0423] FIG. 19 shows how the predefined distance, optical density threshold, and predefined threshold correlated with the response of tissue samples from cancer patients treated with anti-FRα ADC. FIG. 19 shows the correlation between the actual outcomes of PDX studies of 112 patients with ovarian, NSCLC, colon, and endometrial cancer. The tissues analyzed in FIG. 19 were mouse tissues with tumor cells from 112 human cancer patients. A dose of 5 mg / kg was administered in the PDX studies. FIG. 19 shows the relationship between tumor growth rate (e.g., 1.5=150%) and predicted efficacy score. bystander_memb(meanOD)_binary_r25_cut12 (percentage of cancer cells that are (i) optical density (OD) positive or (ii) optical density (OD) negative and within 25 microns of OD positive cells). Y-axis values ​​above 0 indicate that the tumor increased in size during the observation period, and values ​​below 0 indicate that the tumor decreased in size during the observation period.

[0424] FIG. 19 shows that there was a negative correlation between the predicted efficacy score bystander_memb(meanOD)_binary_r25_cut12 and the change in tumor size after treatment. In the bottom plot of FIG. 19, Spearman's rho is −0.57 and p is equal to 5.77E-8. Thus, there was a greater reduction in tumor size associated with a higher predicted efficacy score. The top bar graph of FIG. 19 shows the predicted efficacy scores by decreasing size for each of the 112 PDX models of patients with ovarian cancer, NSCLC, colon cancer, and endometrial cancer.

[0425] Figure 20 is a plot showing predicted efficacy scores by increasing size of 28 PDX models of patients with ovarian cancer, NSCLC, and endometrial cancer. The 98% predefined threshold used in step 16 of method 10 was determined based on 28 PDX models. Circles represent PDX models of ovarian cancer, squares represent PDX models of lung cancer (NSCLC), and triangles represent PDX models of endometrial cancer. Enclosed shapes represent PDX models that were non-responsive to ADC, and shapes without circles represent PDX models that showed a positive response to ADC.

[0426] FIG. 21 shows how patients were stratified to determine the optimal predefined threshold between responsive and non-responsive patients. FIG. 21 is a receiver operating characteristic (ROC) curve used to evaluate the discriminatory performance of a predictor, in this case the predictive validity score. The "R package cutpointr" ROC curve analysis tool was used to determine the optimal predefined threshold, as well as predictive performance parameters such as AUC (area under the curve), Youden index value, sensitivity, and specificity. AUC values ​​close to 1 indicate good discriminatory performance in stratifying patients between binary clinical outcomes such as responsive (R) and non-responsive (NR). Sensitivity is the true positive rate, which is equal to the number of true positives divided by the sum of true positives and false negatives [TP / (TP+FN)]. Sensitivity indicates the probability that a positive score actually corresponds to a positive clinical outcome. Specificity is the false positive rate, which is equal to the number of true negatives divided by the sum of true negatives and false positives [TN / (TN+FP)]. Specificity indicates the probability that a positive score actually corresponds to a negative clinical outcome. The Youden index value is equal to sensitivity + specificity -1.

[0427] For the predictive validity score bystander_memb(meanOD)_binary_r25_cut12, Figure 21 shows that the optimal predefined threshold is 98% with an area under the curve (AUC) of 0.8235. The sensitivity is 0.8824, the specificity is 0.8182 and the Youden index value is 0.7005.

[0428] FIG. 22 shows the distribution of values ​​of the predicted efficacy score bystander_memb(meanOD)_binary_r25_cut12 in 197 tissue samples from a database of ovarian cancer patients. Stained tissue samples from human patients were analyzed in contrast to tissues from PDX models. The bar graph of decreasing score values ​​shows the prevalence among ovarian cancer patients of various ranges of predicted efficacy scores (binary space proximity scores). A quarter of the patients had a score below 60.11%. The average score was 77.69%. FIG. 22 shows that tissue samples from the majority (109) of the 197 ovarian cancer patients yielded a bystander_memb(meanOD)_binary_r25_cut12 score of 98% or higher. These ovarian cancer patients are more likely to have a positive response to therapy involving anti-FRα antibody drug conjugates.

[0429] In step 17 of method 10 of Figure 9, therapy involving an anti-FRα ADC is recommended for the cancer patient if the predicted efficacy score exceeds a predetermined threshold of 98%. Thus, anti-FRα ADC therapy is recommended for 109 of the 197 ovarian cancer patients.

[0430] D. Validation of the Percent Cells with Minimum OD A variation of the novel method of FIG. 9 was also validated by treating tumors in a PDX model with anti-FRα ADC AB1370049-SG3932-DAR8 at doses of 2.5 mg / kg and 5 mg / kg. FIG. 23 is a flow chart of steps 33-38 of method 32 for predicting a cancer patient's response to an anti-FRα ADC. Method 32 is another embodiment of method 10 of FIG. 9, which predicts the efficacy of therapy with an anti-FRα ADC based on the percentage of cancer cells that show a minimal amount of staining. In step 33, a digital image is obtained of a tissue slice from a cancer patient stained with a dye linked to a diagnostic antibody that targets the same protein (in this case, the FRα protein) as is targeted by the ADC therapy. In step 34, image analysis is performed on the digital image to generate image objects of the cancer cells and cell membranes. In step 35, the average optical density (OD) of the staining of the dye in the cell membrane is determined for each cancer cell.

[0431] In step 36, each cancer cell is identified as either optical density positive or optical density negative based on the average optical density of the cell membrane staining. If the average optical density is equal to or greater than the optical density threshold, the cancer cell is identified as optical density positive. If the average optical density is less than the optical density threshold, the cancer cell is optical density negative. In this embodiment, the optical density threshold is 12 on a scale of 220 maximum optical density.

[0432] In step 37, a predicted efficacy score is generated for the tissue sample based on the percentage of cancer cells in the digital image that are optical density positive. If the percentage of cancer cells that are optical density positive is equal to or greater than the percentage threshold, the predicted efficacy score is positive. If the percentage of cancer cells that are optical density positive is less than the percentage threshold, the predicted efficacy score is negative. The optical density threshold and the percentage threshold correlate with the response of training patients treated with the ADC. In this embodiment, the percentage threshold is 90 percent.

[0433] FIG. 24 shows how the optical density threshold and percentage threshold correlated with the response of tissue samples from cancer patients treated with anti-FRα ADC. FIG. 24 shows the correlation between the actual outcomes of a PDX study of 112 patients with ovarian, NSCLC, colon, and endometrial cancer. The tissues analyzed in FIG. 24 were mouse tissues with tumor cells from 112 human cancer patients. A dose of 5 mg / kg was administered in the PDX study. FIG. 24 shows the relationship between tumor growth in percent (1=100%) and the predicted efficacy score memb(meanOD)_perc_pos_12 (the percentage of cancer cells with a mean optical density of staining of at least 12 on a scale of 220). Y-axis values ​​above 0 indicate that the tumor increased in size during the observation period, and values ​​below 0 indicate that the tumor decreased in size during the observation period.

[0434] Figure 24 shows that there was a negative correlation between the predicted efficacy score memb(meanOD)_perc_pos_12 and the change in tumor size after treatment. In the bottom plot of Figure 24, Spearman's rho is -0.54 and p equals 3.36E-7. Thus, there was a greater reduction in tumor size associated with a higher predicted efficacy score. The top bar graph of Figure 24 shows the predicted efficacy scores ordered by decreasing size for each of the 112 PDX models of patients with ovarian cancer, NSCLC, colon cancer, and endometrial cancer.

[0435] Figure 25 is a plot showing predicted efficacy scores ordered by increasing size of 28 PDX models of patients with ovarian cancer, NSCLC, and endometrial cancer. The 90% percentage threshold used in step 37 of method 32 was determined based on 28 PDX models. Circles represent PDX models of ovarian cancer, squares represent PDX models of lung cancer (NSCLC), and triangles represent PDX models of endometrial cancer. Enclosed shapes represent PDX models that were non-responsive to ADC, and shapes without circles represent PDX models that showed a positive response to ADC.

[0436] FIG. 26 shows how patients were stratified to determine the optimal percentage threshold between responsive and non-responsive patients. FIG. 26 is a receiver operating characteristic (ROC) curve used to evaluate the discriminatory performance of a predictor, in this case the memb(meanOD)_perc_pos_12 score. The same ROC curve analysis tool used with FIG. 20 was used to determine the optimal percentage threshold in FIG. 25. For the predictive efficacy score memb(meanOD)_perc_pos_12, FIG. 26 shows that the optimal percentage threshold is 90% with an area under the curve (AUC) of 0.7968. The sensitivity is 0.8235, the specificity is 0.8182, and the Youden index value is 0.6417.

[0437] FIG. 27 shows the distribution of values ​​of the predicted efficacy score memb(meanOD)_perc_pos_12 in 197 tissue samples from a database of ovarian cancer patients. In FIG. 27, stained tissue samples from human patients were analyzed as opposed to tissue from PDX models. The bar graph of decreasing score values ​​shows the prevalence among ovarian cancer patients of different ranges of predicted efficacy scores. A quarter of the patients had a score below 26.08%. The average score was 66.75%. FIG. 27 shows that tissue samples from just under half (96 patients) of the 197 ovarian cancer patients yielded a memb(meanOD)_perc_pos_12 score of 90% or higher. These ovarian cancer patients are more likely to have a positive response to therapy with anti-FRα antibody drug conjugates.

[0438] In step 38 of method 32 of Figure 23, if the predicted efficacy score is positive, therapy with anti-FRα ADC is recommended for the cancer patient. The score is positive if the memb(meanOD)_perc_pos_12 score is 90% or higher. Thus, anti-FRα ADC therapy is recommended for just under half of the 197 ovarian cancer patients.

[0439] Verification of the median of absolute deviation of E.OD Another embodiment of method 10 of FIG. 9 involves determining the median absolute deviation of staining for each cell from the median membrane staining of all cancer cells in the digital image. FIG. 28 is a flow chart of steps 41-47 of method 40 for predicting a cancer patient's response to an anti-FRα ADC. Method 40 predicts the efficacy of a therapy involving an anti-FRα ADC based on the absolute deviation of membrane staining in the digital image. In step 41, a digital image of a tissue slice from a cancer patient stained with a dye linked to a diagnostic antibody targeting the FRα protein is obtained. In step 42, image analysis is performed on the digital image to generate image objects of the cancer cells and cell membranes. In step 43, the average optical density (OD) of the staining of the dye in the cell membrane is determined for each cancer cell. In step 44, the median optical density of all cancer cells in the digital image is determined.

[0440] In step 45, the median absolute deviation of the optical density of the cancer cells is determined from the median optical density of all cancer cells in the digital image. In step 46, a predicted efficacy score is generated for the tissue sample based on the median absolute deviation. The predicted efficacy score is positive if the median absolute deviation is equal to or greater than a deviation threshold and negative if the median absolute deviation is less than a deviation threshold. The deviation threshold correlates with the response of a cohort of training patients treated with the ADC. In this embodiment, the deviation threshold is 24%.

[0441] FIG. 29 shows how the optical density threshold and deviation threshold correlated with the response of tissue samples from cancer patients treated with anti-FRα ADC. FIG. 29 shows the correlation between the actual outcomes of a PDX study of 112 patients with ovarian, NSCLC, colon, and endometrial cancer. The tissues analyzed in FIG. 29 were mouse tissues with tumor cells from 112 human cancer patients. A dose of 5 mg / kg was administered in the PDX study. FIG. 29 shows the relationship between tumor growth in percent (1=100%) and the predicted efficacy score hist_memb(meanOD)_mad (median absolute deviation of optical density of cancer cells from the median optical density of all cancer cells in the digital image). Y-axis values ​​above 0 indicate that the tumor size increased during the observation period, and values ​​below 0 indicate that the tumor size decreased during the observation period.

[0442] Figure 29 shows that there was a negative correlation between the predicted efficacy score hist_memb(meanOD)_mad and the change in tumor size after treatment. In the bottom plot of Figure 29, Spearman's rho is -0.60 and p is equal to 7.03E-9. Thus, there was a greater reduction in tumor size associated with a higher predicted efficacy score. The top bar graph of Figure 29 shows the predicted efficacy scores ordered by decreasing size for each of the 112 PDX models of patients with ovarian cancer, NSCLC, colon cancer, and endometrial cancer.

[0443] Figure 30 is a plot showing predicted efficacy scores ordered by increasing size for 28 PDX models of patients with ovarian cancer, NSCLC, and endometrial cancer. The 24% deviation threshold used in step 46 of method 40 was determined based on 28 PDX models. Circles represent PDX models of ovarian cancer, squares represent PDX models of lung cancer (NSCLC), and triangles represent PDX models of endometrial cancer. Enclosed shapes represent PDX models that were non-responsive to ADC, and shapes without circles represent PDX models that showed a positive response to ADC.

[0444] FIG. 31 shows how patients were stratified to determine the optimal deviation threshold between responsive and non-responsive patients. FIG. 31 is a receiver operating characteristic (ROC) curve used to evaluate the discriminatory performance of a predictor, in this case, the hist_memb(meanOD)_mad score. The same ROC curve analysis tool used with FIG. 20 was used to determine the optimal deviation threshold in FIG. 30. For the predictive efficacy score hist_memb(meanOD)_mad, FIG. 31 shows that the optimal deviation threshold is 24% with an area under the curve (AUC) of 0.8128. The sensitivity is 0.7059, the specificity is 0.9091, and the Youden index value is 0.6150.

[0445] FIG. 32 shows the distribution of values ​​of the predicted efficacy score hist_memb(meanOD)_mad in 197 tissue samples from a database of ovarian cancer patients. In FIG. 32, stained tissue samples from human patients were analyzed, in contrast to tissues from PDX models shown in FIG. 30. The bar graphs with decreasing score values ​​show the prevalence among ovarian cancer patients of various ranges of predicted efficacy scores. A quarter of the patients had a score of 3.20% or less. A quarter of the patients had a score of 44.59% or more. The average score was 25.90%. FIG. 32 shows that tissue samples from just over half (103 patients) of the 197 ovarian cancer patients yielded a hist_memb(meanOD)_mad score of 24% or more. These ovarian cancer patients are more likely to have a positive response to therapy involving anti-FRα antibody drug conjugates.

[0446] In step 47 of method 40 of Figure 28, if the predicted efficacy score is positive, therapy with anti-FRα ADC is recommended for the cancer patient. If the hist_memb(meanOD)_mad score is 24% or higher, the score is positive. Thus, anti-FRα ADC therapy is recommended for just over half of the 197 ovarian cancer patients.

[0447] F. Validation of Mean OD to Predict Dosing Efficacy Another embodiment of method 10 of FIG. 9 involves determining the median optical density of all cancer cells in a digital image to determine a recommended dose of anti-FRα ADC. FIG. 33 is a flow chart of steps 49-54 of method 48 for predicting the efficacy of various doses of anti-FRα ADC based on the median optical density of membrane staining in a digital image. In step 49, a digital image of a tissue slice from a cancer patient stained with a dye linked to a diagnostic antibody targeting the FRα protein is obtained. In step 50, image analysis is performed on the digital image to generate image objects of cancer cells and cell membranes. In step 51, the average optical density (OD) of the staining of the dye in the cell membrane is determined for each cancer cell. In step 52, the median optical density of all cancer cells in the digital image is determined. This is equivalent to determining the optical density of the 50th percentile (q50) of the cancer cells in the digital image, which is the whole slide image (WSI).

[0448] In step 53, the recommended dose of the anti-FRα ADC is determined based on whether (i) the median optical density is below the lower optical density threshold, (ii) between the lower and upper optical density thresholds, or (iii) above the upper optical density threshold. If the median optical density is below the lower optical density threshold, the recommended dose is zero. Thus, the ADC is not administered to the cancer patient if the median optical density is below the lower optical density threshold. If the median optical density is between the lower and upper optical density thresholds, the recommended dose is the higher dose. If the median optical density is above the upper optical density threshold, the recommended dose is the lower dose. The lower and upper optical density thresholds correlate with the response of a cohort of training patients treated with the ADC. In this embodiment, the lower dose is 2.5 mg / kg and the higher dose is 5 mg / kg. In this embodiment, the lower optical density threshold is 25 and the upper optical density threshold is 39, both on a maximum optical density scale of 220.

[0449] Figures 34-39 show how the lower and upper optical density thresholds correlated with the response of tissue samples from cancer patients treated with anti-FRα ADC. Figure 34 shows the correlation between the actual outcomes of a PDX study of 28 patients with ovarian, NSCLC, and endometrial cancer. The tissues analyzed in Figure 34 were mouse tissues with tumor cells from 28 human cancer patients. A dose of 5 mg / kg was administered in the PDX study. Figure 34 shows the relationship between tumor growth in percent and the predicted efficacy score dual_dose_memb(meanOD)_q50 (mean OD of membrane staining). Y-axis values ​​above 0 indicate that the tumor increased in size during the observation period, while values ​​below 0 indicate that the tumor decreased in size during the observation period.

[0450] Figure 34 shows that there was a negative correlation between the predicted efficacy score dual_dose_memb(meanOD)_q50 and the change in tumor size after treatment. There was a greater reduction in tumor size associated with higher predicted efficacy scores.

[0451] Figure 35 is a plot showing predicted efficacy scores ordered by increasing size for 28 PDX models of patients with ovarian cancer, NSCLC, and endometrial cancer shown in Figure 34. The 25 lower optical density threshold used in step 53 of method 48 was determined based on 28 PDX models treated with 5 mg / kg anti-FRα ADC. Circles represent PDX models of ovarian cancer, squares represent PDX models of lung cancer (NSCLC), and triangles represent PDX models of endometrial cancer. Enclosed shapes represent PDX models that were non-responsive to ADC, and shapes without circles represent PDX models that showed a positive response to ADC.

[0452] FIG. 36 shows how patients were stratified to determine the optimal lower optical density threshold between responding and non-responding patients. FIG. 36 is a receiver operating characteristic (ROC) curve used to evaluate the discriminatory performance of a predictor, in this case, dual_dose_memb(meanOD)_q50 score. For predicting efficacy scores for PDX models treated with 5 mg / kg anti-FRα ADC, FIG. 36 shows that the optimal lower optical density threshold is an optical density of 25, with an area under the curve (AUC) of 0.7914. The sensitivity is 0.8824, the specificity is 0.7273, and the Youden index value is 0.6096.

[0453] FIG. 37 shows how the upper optical density threshold correlated with the response of tissue samples from cancer patients treated with anti-FRα ADC. FIG. 37 shows the correlation between actual outcomes of PDX studies of 23 patients with ovarian, NSCLC, and endometrial cancer. A dose of 2.5 mg / kg was administered in the PDX studies. The tissues analyzed in FIG. 37 were mouse tissues bearing tumor cells from 23 human cancer patients. FIG. 37 shows the relationship between predicted efficacy score dual_dose_memb(meanOD)_q50 and percent tumor growth after administration of a dose of 2.5 mg / kg. FIG. 37 shows that there was a negative correlation between predicted efficacy score dual_dose_memb(meanOD)_q50 and change in tumor size after treatment. There was a greater reduction in tumor size associated with a higher predicted efficacy score.

[0454] Figure 38 is a plot showing predicted efficacy scores ordered by increasing size for 23 PDX models of patients with ovarian cancer, NSCLC, and endometrial cancer shown in Figure 37. The 39 upper optical density thresholds used in step 53 of method 48 were determined based on 23 PDX models treated with 2.5 mg / kg anti-FRα ADC.

[0455] Figure 39 shows how patients were stratified to determine the optimal upper optical density threshold between responding and non-responding patients. Figure 39 is the ROC curve used to evaluate the discrimination performance of the dual_dose_memb(meanOD)_q50 predicted efficacy score. For the predicted efficacy score of the PDX model treated with 2.5 mg / kg anti-FRα ADC, Figure 39 shows that the optimal upper optical density threshold is an optical density of 39 with an area under the curve (AUC) of 0.8730. The sensitivity is 0.7778, the specificity is 0.9286, and the Youden index value is 0.7063.

[0456] Figure 40 is a graph of the median optical density of each of the PDX models compared to the median optical density of human ovarian tumors. One curve shows the median optical density of membrane staining for 197 human ovarian tumors, with a dashed line showing the average of all median optical density scores. The other curve shows the median optical density of membrane staining for 18 PDX models, with a dashed line showing the average of the median optical density scores, slightly to the left of the dashed line for human ovarian tumors. The graph demonstrates that the distribution of FRα protein in ovarian cancer PDX models is comparable to that in profiled human ovarian tumors, and there is a similar distribution of FRα in patient-derived xenografts and human tumors of ovarian cancer. Thus, PDX models can be used to determine various predictive efficacy score thresholds.

[0457] FIG. 41 shows the distribution of values ​​of the predicted efficacy score dual_dose_memb(meanOD)_q50 in 197 tissue samples from a database of ovarian cancer patients. In FIG. 41, stained tissue samples from human patients were analyzed, as opposed to tissues from PDX models shown in FIGS. 34-39. The bar graphs of decreasing score values ​​show the prevalence among ovarian cancer patients of various ranges of predicted efficacy scores. A quarter of the patients had an optical density score below 8.74, and another quarter of the patients had an optical density score above 99.28. The mean score was an optical density of 56.21. FIG. 41 shows that tissue samples from 101 of the 197 ovarian cancer patients yielded a dual_dose_memb(meanOD)_q50 score of 39 or higher. These ovarian cancer patients would be more likely to have a positive response to the lower dose of 2.5 mg / kg of anti-FRα antibody drug conjugate. An additional 8 ovarian cancer patients would likely have a positive response to the higher dose of 5 mg / kg of anti-FRα antibody drug conjugate. These additional 8 patients would likely not achieve a positive response only at the lower dose of 2.5 mg / kg. 101 patients would also likely have a positive response at the higher 5 mg / kg dose of ADC, but these patients would be able to avoid any adverse side effects of the additional ADC dose. This is not necessary for these patients to achieve a positive response. The recommended dose is zero for the 88 patients with a predicted efficacy score of less than 25. Therapy with anti-FRα ADC would not be recommended for the 88 patients.

[0458] In step 54 of method 48 of FIG. 33, therapy with an anti-FRα ADC at a recommended dosage is recommended to a cancer patient based on whether a predicted efficacy score for a digital image of the patient's tissue sample is below a lower optical density threshold of 25, between a lower optical density threshold of 25 and an upper optical density threshold of 39, or equal to or greater than an upper optical density threshold of 39. A higher dosage of 5 mg / kg is recommended if the predicted efficacy score is between 25 and 39, and a lower dosage of 2.5 mg / kg is recommended if the predicted efficacy score is above 39. If the predicted efficacy score is less than 25, the recommended dosage is zero.

[0459] G. Verification of the difference between membrane OD and cytoplasmic OD Yet another embodiment of the novel method for generating a predictive efficacy score involves determining the difference between the optical density of the membrane staining and the optical density of the cytoplasmic staining. FIG. 42 is a flow chart of steps 56-62 of a method 55 for predicting a cancer patient's response to an anti-FRα ADC. The method 55 predicts the efficacy of a therapy involving an anti-FRα ADC by identifying the 85% quantile of the difference between the optical density of the membrane staining and the optical density of the cytoplasmic staining. In step 56, a digital image of a tissue slice from a cancer patient stained with a dye linked to a diagnostic antibody targeting the FRα protein is obtained. In step 57, image analysis is performed on the digital image to generate image objects of the cancer cells, cytoplasm, and cell membrane. In step 58, the average optical density (OD) of the staining of the dye in the cell membrane is determined for each cancer cell. In step 59, the average optical density of the staining of the dye in the cytoplasm is determined for each cancer cell. In step 60, the difference between the average optical density of the membrane staining and the average optical density of the cytoplasmic staining in the digital image is determined for each cancer cell. In step 61, the 85% quantile of the difference between the optical density of the membrane staining and the optical density of the cytoplasmic staining in the digital image is identified among all cancer cells. In step 62, if the 85% quantile of the optical density difference exceeds a predetermined difference threshold, therapy with anti-FRα ADC is recommended for the cancer patient. The difference threshold correlates with the response of a cohort of training patients treated with the ADC. In this embodiment, the difference threshold is 23.

[0460] FIG. 43 shows how the difference threshold correlated with the response of tissue samples from cancer patients treated with anti-FRα ADC. FIG. 43 shows the correlation between the actual outcomes of PDX studies of patients with ovarian, NSCLC, colon, and endometrial cancer. The tissues analyzed in FIG. 43 were mouse tissues with tumor cells from human cancer patients. In the PDX studies, a dose of 5 mg / kg was administered. FIG. 43 shows the relationship between tumor growth in percent (e.g., 2.5=250%) and the predicted efficacy score memb(meanOD)-cyto(meanOD)_q85 (85% quantile of the difference between the OD of membrane staining and the OD of cytoplasmic staining). Y-axis values ​​above 0 indicate that the tumor size increased during the observation period, and values ​​below 0 indicate that the tumor size decreased during the observation period. FIG. 43 shows that there was a negative correlation between the predicted efficacy score memb(meanOD)-cyto(meanOD)_q85 and the change in tumor size after treatment. In the plot of Figure 43, Spearman's rho is -0.62, p is equal to 1.50E-9. Thus, there was a greater reduction in tumor size associated with a higher predictive efficacy score.

[0461] Figure 44 is a plot showing the predicted efficacy scores ordered by increasing size of 28 PDX models of patients with ovarian cancer, NSCLC, and endometrial cancer. The difference threshold of 23 used in step 62 of method 55 was determined based on 28 PDX models. Circles represent PDX models of ovarian cancer, squares represent PDX models of lung cancer (NSCLC), and triangles represent PDX models of endometrial cancer. Enclosed shapes represent PDX models that were non-responsive to ADC, and shapes without circles represent PDX models that showed a positive response to ADC.

[0462] Figure 45 shows how patients were stratified to determine the optimal difference threshold between responsive and non-responsive patients. Figure 45 is the ROC curve used to evaluate how well the predictive efficacy score memb(meanOD)-cyto(meanOD)_q85 discriminated between responsive and non-responsive patients. Figure 45 shows that the optimal difference threshold is 23, with an area under the curve (AUC) of 0.8289. The sensitivity is 0.7647, the specificity is 0.9091, and the Youden index value is 0.6738.

[0463] Figure 46 shows the distribution of values ​​of the predicted efficacy score memb(meanOD)-cyto(meanOD)_q85 in 197 tissue samples from a database of ovarian cancer patients. In Figure 46, stained tissue samples from human patients were analyzed, in contrast to tissues from PDX models shown in Figure 44. The bar graphs with decreasing score values ​​show the prevalence among ovarian cancer patients of various ranges of predicted efficacy scores. A quarter of the patients had a score of 5.01 or less. A quarter of the patients had a score of 41.85 or more. The average score was 24.48. Figure 46 shows that tissue samples from just over half (104 patients) of the 197 ovarian cancer patients yielded a memb(meanOD)-cyto(meanOD)_q85 score of 23 or more. These ovarian cancer patients are more likely to have a positive response to therapy involving anti-FRα antibody drug conjugates.

[0464] In step 62 of method 55 of Figure 42, therapy with an anti-FRα ADC is recommended for a cancer patient if the 85% quantile of the difference between membrane optical density and cytoplasmic optical density for the patient's tissue sample exceeds a predetermined difference threshold. In this example, the predetermined difference threshold is 23. Thus, anti-FRα ADC therapy is recommended for 104 of the 197 ovarian cancer patients.

[0465] Although the present invention has been described in connection with certain specific embodiments for instructional purposes, the invention is not limited thereto, and various modifications, adaptations, and combinations of various features of the described embodiments may be made without departing from the scope of the invention as set forth in the claims.

[0466] XII. Preferred Embodiments of the Invention Preferred embodiments of the invention are described below and are referred to as embodiments E1 to E32.

[0467] E1. A method of generating a recommended dosage for treating a cancer patient with an antibody drug conjugate (ADC), comprising an ADC payload and an ADC antibody or antigen-binding fragment thereof that targets a protein on a cancer cell, the protein being folate receptor alpha (FRα), the method comprising:

[0468] immunohistochemically staining a tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample;

[0469] acquiring a digital image of the tissue sample;

[0470] detecting cancer cells in the digital image; and

[0471] calculating for each cancer cell the average optical density of staining with the dye in the membrane of said cancer cell;

[0472] calculating the median optical density of all cancer cells in said digital image;

[0473] determining the recommended dosage based on whether the median optical density is below a lower optical density threshold, between the lower and upper optical density thresholds, or above the upper optical density threshold, wherein the recommended dosage is zero if the median optical density is below the lower optical density threshold, a higher dosage if the median optical density is between the lower and upper optical density thresholds, and a lower dosage if the median optical density is above the upper optical density threshold, wherein the lower optical density threshold and the upper optical density threshold correlate with the response of a cohort of training patients treated with the ADC;

[0474] and recommending to the cancer patient a therapy involving the ADC and the recommended dosage.

[0475] E2.

[0476] The method of E1, further comprising administering to the cancer patient the recommended dose of the therapy involving the ADC.

[0477] E3. The method of E1, wherein the lower optical density threshold is in the range of 21 to 29 on a scale having a maximum optical density of 220.

[0478] E4. The method of any of E1-E3, wherein the upper optical density threshold is in the range of 35 to 43 on a scale having a maximum optical density of 220.

[0479] E5. The method according to any one of E1 to E4, wherein the lower dose is 2.5 mg / kg or less and the higher dose is 5 mg / kg or more.

[0480] E6. A method of generating a response score to predict a cancer patient's response to an antibody drug conjugate (ADC) comprising an ADC payload and an ADC antibody or antigen-binding fragment thereof that targets a protein on a cancer cell, the protein being folate receptor alpha (FRα), the method comprising:

[0481] immunohistochemically staining a tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample;

[0482] acquiring a digital image of the tissue sample;

[0483] detecting cancer cells in the digital image; and

[0484] calculating for each cancer cell the average optical density of staining with the dye in the membrane of said cancer cell;

[0485] calculating the median optical density of all cancer cells in said digital image;

[0486] generating a response score for the tissue sample based on the median optical density, the response score being positive if the median optical density is equal to or greater than an optical density threshold and negative if the median optical density is less than the optical density threshold, the optical density threshold correlating with the response of a cohort of training patients treated with the ADC;

[0487] and if the response score is positive, recommending therapy involving the ADC to the cancer patient.

[0488] E7. A method of generating a response score to predict a cancer patient's response to an antibody drug conjugate (ADC) comprising an ADC payload and an ADC antibody that targets a protein on a cancer cell, the protein being folate receptor alpha (FRα), the method comprising:

[0489] immunohistochemically staining a tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample;

[0490] acquiring a digital image of the tissue sample;

[0491] detecting cancer cells in the digital image; and

[0492] determining for each cancer cell the optical density of staining of the dye in the membrane of the cancer cell;

[0493] identifying each cancer cell as either optical density positive if the optical density of the cancer cell is equal to or greater than an optical density threshold, or optical density negative if the mean optical density of the cancer cells is less than the optical density threshold;

[0494] generating a response score for the tissue sample based on the percentage of cancer cells in the digital image that are optical density positive, the response score being positive if the percentage of the cancer cells that are optical density positive is equal to or greater than a percentage threshold and being negative if the percentage of the cancer cells that are optical density positive is less than the percentage threshold, the optical density threshold and the percentage threshold correlating with the response of a cohort of training patients treated with the ADC;

[0495] and if the response score is positive, recommending therapy involving the ADC to the cancer patient.

[0496] E8.

[0497] The method of E6 or E7, further comprising administering to the cancer patient the therapy involving the ADC if the response score is positive.

[0498] E9. The method according to any one of E6 to E8, wherein the percentage threshold is in the range of 80% to 90%.

[0499] E10. A method of identifying a cancer patient for treatment with an antibody drug conjugate (ADC), comprising an ADC payload and an ADC antibody that targets a protein on a cancer cell, said protein being folate receptor alpha (FRα), said method comprising:

[0500] immunohistochemically staining a tissue sample from the cancer patient using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample;

[0501] acquiring a digital image of the tissue sample;

[0502] detecting cancer cells in the digital image; and

[0503] calculating for each cancer cell the average optical density of staining with the dye in the membrane of said cancer cell;

[0504] identifying each cancer cell as either optical density positive if the mean optical density of the cancer cells is equal to or greater than an optical density threshold, or optical density negative if the mean optical density of the cancer cells is less than the optical density threshold;

[0505] generating a proximity score for the tissue sample equal to the percentage of cancer cells in the digital image that are either optical density positive or optical density negative, but within a predetermined distance from optical density positive cancer cells;

[0506] and if the proximity score exceeds a predetermined percentage threshold, identifying the cancer patient as one who would likely benefit from administration of the ADC.

[0507] E11.

[0508] The method of E10, further comprising recommending a therapy involving the ADC to the cancer patient if the proximity score exceeds the predetermined percentage threshold.

[0509] E12.

[0510] The method of E10 or E11, further comprising administering to said cancer patient said therapy involving said ADC if said proximity score exceeds said predetermined percentage threshold.

[0511] E13. The method according to any one of E10 to E12, wherein the percentage threshold is in the range of 95% to 100%.

[0512] E14. The method according to any of E7 to E13, wherein the optical density threshold is in the range of 10 to 15 on a scale having a maximum optical density of 220.

[0513] E15. A method of generating a response score to predict a cancer patient's response to an antibody drug conjugate (ADC), comprising an ADC payload and an ADC antibody that targets a protein on a cancer cell, the protein being folate receptor alpha (FRα), the method comprising:

[0514] immunohistochemically staining a tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample;

[0515] acquiring a digital image of the tissue sample;

[0516] detecting cancer cells in the digital image; and

[0517] calculating for each cancer cell the average optical density of staining with the dye in the membrane of said cancer cell;

[0518] calculating the median optical density of all cancer cells in said digital image;

[0519] calculating a median absolute deviation of the optical density of the cancer cells from the median of the optical density of all cancer cells in the digital image;

[0520] generating a response score for the tissue sample based on the median absolute deviation, the response score being positive if the median absolute deviation is equal to or greater than a deviation threshold and negative if the median absolute deviation is less than the deviation threshold, the deviation threshold correlating with the response of a cohort of training patients treated with the ADC;

[0521] and if the response score is positive, recommending therapy involving the ADC to the cancer patient.

[0522] E16.

[0523] The method of E15, further comprising administering to the cancer patient the therapy involving the ADC if the predicted efficacy score is positive.

[0524] E17. The method according to E15 or E16, wherein the deviation threshold is in the range of 10 to 15 on an optical density scale having a maximum value of 220.

[0525] E18. The method according to any of E7 to E17, wherein said detection of cancer cells involves detecting, for each cancer cell, pixels belonging to said membrane using the cell centroid determined for each cancer cell.

[0526] E19. The method according to any of E1-E18, wherein the staining intensity of each membrane is calculated based on the average optical density of the brown diaminobenzidine (DAB) signal in the pixels of said membrane.

[0527] E20. The method of any of E1 to E19, wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

[0528] E21. The method according to any one of E1 to E20, wherein the dye is 3,3'-diaminobenzidine (DAB).

[0529] E22. The method of any of E1-E21, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer, and malignant pleural mesothelioma.

[0530] E23. The method according to any of E1-E22, wherein said cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

[0531] E24. The method of E23, wherein said NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous and adenocarcinoma NSCLC.

[0532] E25. The method of any of E1-E24, wherein the ADC payload is a cytotoxin.

[0533] E26. The method of E25, wherein said cytotoxin is a topoisomerase I inhibitor.

[0534] E27. The topoisomerase I inhibitor has the following formula:

[0535] [ka]

[0536] and salts and solvates thereof, wherein R represents a linker attached to said ADC antibody or antigen-binding fragment thereof.

[0537] E28. The ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is:

[0538] [ka]

[0539] [ka]

[0540] [ka]

[0541] [ka] The method according to any one of E1 to E27, selected from the group consisting of

[0542] E29. The ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is:

[0543] [ka] The method according to any one of E1 to E28,

[0544] E30. The method of any of E1-E29, wherein the ADC has a drug-to-antibody ratio (DAR) within a range selected from the group consisting of 1-20, 1-10, 2-10, 2-8, 2-6, and 4-10.

[0545] The method of any of E1-E30, wherein said ADC has a drug-to-antibody ratio (DAR) selected from the group consisting of 4 and 8.

[0546] E32. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of:

[0547] (a) heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, light chain CDR2 (KASGLES) of SEQ ID NO: 5, and light chain CDR3 (QQYNSYSQLT) of SEQ ID NO: 6;

[0548] (b) a heavy chain CDR1 (SYAMS) of SEQ ID NO: 7, a heavy chain CDR2 (SISSGRSYIYYADSVKG) of SEQ ID NO: 8, a heavy chain CDR3 (EMQQLALDY) of SEQ ID NO: 9, a light chain CDR1 (RASQGISNFLA) of SEQ ID NO: 10, a light chain CDR2 (AASSLQS) of SEQ ID NO: 11, and a light chain CDR3 (QQYNSYPFT) of SEQ ID NO: 12;

[0549] (c) a heavy chain CDR1 (SNSAAWN) of SEQ ID NO: 13, a heavy chain CDR2 (RTYYRSNWYNDYTLSVKS) of SEQ ID NO: 14, a heavy chain CDR3 (GVGRFDS) of SEQ ID NO: 15, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 16, a light chain CDR2 (KASSLES) of SEQ ID NO: 17, and a light chain CDR3 (QEYKTYSIFT) of SEQ ID NO: 18;

[0550] (d) heavy chain CDR1 (SYNMN) of SEQ ID NO: 19, heavy chain CDR2 (SISSGSSYIYYADSMKG) of SEQ ID NO: 20, heavy chain CDR3 (GMTTLTFDY) of SEQ ID NO: 21, light chain CDR1 (RASQGISTFLA) of SEQ ID NO: 22, light chain CDR2 (AASSLQS) of SEQ ID NO: 23, and light chain CDR3 (QQYISYPLT) of SEQ ID NO: 24;

[0551] (e) heavy chain CDR1 (SYSMN) of SEQ ID NO: 25, heavy chain CDR2 (SISSRSSYVYYADSVKG) of SEQ ID NO: 26, heavy chain CDR3 (GMTTLTFDY) of SEQ ID NO: 27, light chain CDR1 (RASQGISSFLA) of SEQ ID NO: 28, light chain CDR2 (AASSLQS) of SEQ ID NO: 29, and light chain CDR3 (QQYNSYPLT) of SEQ ID NO: 30, and

[0552] (f) A method described in any of E1 to E31, comprising a plurality of amino acid sequences selected from heavy chain CDR1 (SDSATWN) of SEQ ID NO: 31, heavy chain CDR2 (RTYYRSKWYSDYAVSVKS) of SEQ ID NO: 32, heavy chain CDR3 (GGAPFDY) of SEQ ID NO: 33, light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 34, light chain CDR2 (KASSLES) of SEQ ID NO: 35, and light chain CDR3 (QQYNSYSMYT) of SEQ ID NO: 36.

[0553] E33. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of:

[0554] (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37, and a VL chain comprising the amino acid sequence of SEQ ID NO: 38;

[0555] (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39, and a VL chain comprising the amino acid sequence of SEQ ID NO: 40;

[0556] (c) a VH chain comprising the amino acid sequence of SEQ ID NO: 41, and a VL chain comprising the amino acid sequence of SEQ ID NO: 42;

[0557] (d) a VH chain comprising the amino acid sequence of SEQ ID NO: 43, and a VL chain comprising the amino acid sequence of SEQ ID NO: 44;

[0558] (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45, and a VL chain comprising the amino acid sequence of SEQ ID NO: 46, and

[0559] (f) a VH chain comprising the amino acid sequence of SEQ ID NO: 47, and a VL chain comprising the amino acid sequence of SEQ ID NO: 48.

[0560] E34. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of:

[0561] (a) a heavy chain amino acid sequence of SEQ ID NO: 49, and a light chain amino acid sequence of SEQ ID NO: 50;

[0562] (b) a heavy chain amino acid sequence of SEQ ID NO: 51, and a light chain amino acid sequence of SEQ ID NO: 52;

[0563] (c) a heavy chain amino acid sequence of SEQ ID NO: 53, and a light chain amino acid sequence of SEQ ID NO: 54;

[0564] (d) a heavy chain amino acid sequence of SEQ ID NO: 55, and a light chain amino acid sequence of SEQ ID NO: 56;

[0565] (e) a heavy chain amino acid sequence of SEQ ID NO: 57, and a light chain amino acid sequence of SEQ ID NO: 58, and

[0566] (f) The method described in any one of E1 to E33, comprising a plurality of amino acid sequences selected from the heavy chain amino acid sequence of SEQ ID NO: 59 and the light chain amino acid sequence of SEQ ID NO: 60.

[0567] E35.

[0568] the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, a heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, a heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, a light chain CDR2 (KASGLES) of SEQ ID NO: 5, and a light chain CDR3 (QQYNSYSQLT) of SEQ ID NO: 6;

[0569] the ADC payload is a topoisomerase I inhibitor

[0570] [ka] and

[0571] The method of any of E1 to E34, wherein the ADC has a drug-to-antibody ratio (DAR) of 8.

[0572] E36. The method of E35, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO:37 and a VL chain comprising the amino acid sequence of SEQ ID NO:38.

Claims

1. 1. A method for determining a recommended dosage for treating a cancer patient with an antibody drug conjugate (ADC), comprising an ADC payload and an ADC antibody or antigen-binding fragment thereof that targets a protein on a cancer cell, the protein being folate receptor alpha (FRα), the method comprising: immunohistochemically staining a tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample; acquiring a digital image of the tissue sample; detecting cancer cells in the digital image; and determining for each cancer cell the average optical density of staining with the dye in the membrane of said cancer cell; determining a median optical density of all cancer cells in said digital image; determining the recommended dosage based on whether the median optical density is below a lower optical density threshold, between the lower and upper optical density thresholds, or above the upper optical density threshold, wherein the recommended dosage is zero if the median optical density is below the lower optical density threshold, a higher dosage if the median optical density is between the lower and upper optical density thresholds, and a lower dosage if the median optical density is above the upper optical density threshold, wherein the lower optical density threshold and the upper optical density threshold correlate with the response of a cohort of training patients treated with the ADC; and recommending to the cancer patient a therapy involving the ADC and the recommended dosage.

2. 13. The method of claim 1, further comprising administering to the cancer patient the recommended dosage of the therapy involving the ADC.

3. 2. The method of claim 1, wherein the lower optical density threshold is in the range of 21 to 29 on a scale having a maximum optical density of 220.

4. 2. The method of claim 1, wherein the upper optical density threshold is in the range of 35 to 43 on a scale having a maximum optical density of 220.

5. 2. The method of claim 1, wherein the lower dose is in the range of 2.4 to 2.6 mg / kg and the higher dose is in the range of 4.9 to 5.1 mg / kg.

6. The method of claim 1 , wherein the detection of cancer cells involves detecting, for each cancer cell, pixels belonging to the membrane using a cell centroid determined for each cancer cell.

7. 2. The method of claim 1, wherein the staining intensity of each membrane is calculated based on the average optical density of the brown diaminobenzidine (DAB) signal in the pixels of the membrane.

8. The method of claim 1 , wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

9. The method of claim 1, wherein the dye is 3,3'-diaminobenzidine (DAB).

10. 2. The method of claim 1, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer, and malignant pleural mesothelioma.

11. 10. The method of claim 1, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

12. 12. The method of claim 11, wherein the NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous and adenocarcinoma NSCLC.

13. The method of claim 1 , wherein the ADC payload is a cytotoxin.

14. The method of claim 13, wherein the cytotoxin is a topoisomerase I inhibitor.

15. The topoisomerase I inhibitor has the following formula: 【Chemistry 1】 and salts and solvates thereof, wherein R Ｌ represents a linker connected to the ADC antibody or antigen-binding fragment thereof.

16. The ADC is an anti-FRα antibody, or antigen-binding fragment thereof, conjugated to a topoisomerase I inhibitor, the topoisomerase I inhibitor being 【Chemistry 2】 The method of claim 1 , wherein the compound is selected from the group consisting of

17. The ADC is an anti-FRα antibody, or antigen-binding fragment thereof, conjugated to a topoisomerase I inhibitor, the topoisomerase I inhibitor being 【Chemistry 3】 The method of claim 1, wherein

18. 2. The method of claim 1, wherein the ADC has a drug-to-antibody ratio (DAR) within a range selected from the group consisting of 1-20, 1-10, 2-10, 2-8, 2-6, and 4-10.

19. 2. The method of claim 1, wherein the ADC has a drug-to-antibody ratio (DAR) selected from the group consisting of 4 and 8.

20. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, a heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, a heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, a light chain CDR2 (KASGLES) of SEQ ID NO: 5, and a light chain CDR3 (QQYNSYSSQLT) of SEQ ID NO: 6; (b) a heavy chain CDR1 (SYAMS) of SEQ ID NO: 7, a heavy chain CDR2 (SISSGRSYIYYADSVKG) of SEQ ID NO: 8, a heavy chain CDR3 (EMQQLALDY) of SEQ ID NO: 9, a light chain CDR1 (RASQGISNFLA) of SEQ ID NO: 10, a light chain CDR2 (AASSLQS) of SEQ ID NO: 11, and a light chain CDR3 (QQYNSYPFT) of SEQ ID NO: 12; (c) a heavy chain CDR1 (SNSAAWN) of SEQ ID NO: 13, a heavy chain CDR2 (RTYYRSNWYNDYTLSVKS) of SEQ ID NO: 14, a heavy chain CDR3 (GVGRFDS) of SEQ ID NO: 15, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 16, a light chain CDR2 (KASSLES) of SEQ ID NO: 17, and a light chain CDR3 (QEYKTYSIFT) of SEQ ID NO: 18; (d) a heavy chain CDR1 (SYNMN) of SEQ ID NO: 19, a heavy chain CDR2 (SISSSGSSYIYYADSMKG) of SEQ ID NO: 20, a heavy chain CDR3 (GMTTLTFDY) of SEQ ID NO: 21, a light chain CDR1 (RASQGISTFLA) of SEQ ID NO: 22, a light chain CDR2 (AASSLQS) of SEQ ID NO: 23, and a light chain CDR3 (QQYISYPLT) of SEQ ID NO: 24; (e) a heavy chain CDR1 of SEQ ID NO: 25 (SYSMN), a heavy chain CDR2 of SEQ ID NO: 26 (SISSRSSYVYYADSVKG), a heavy chain CDR3 of SEQ ID NO: 27 (GMTTLTFDY), a light chain CDR1 of SEQ ID NO: 28 (RASQGISSFLA), a light chain CDR2 of SEQ ID NO: 29 (AASSLQS), and a light chain CDR3 of SEQ ID NO: 30 (QQYNSYPLT), and (f) the method of claim 1, comprising a plurality of amino acid sequences selected from heavy chain CDR1 of SEQ ID NO: 31 (SDSATWN), heavy chain CDR2 of SEQ ID NO: 32 (RTYYRSKWYSDYAVSVKS), heavy chain CDR3 of SEQ ID NO: 33 (GGAPFDY), light chain CDR1 of SEQ ID NO: 34 (RASQSISSWLA), light chain CDR2 of SEQ ID NO: 35 (KASSLES), and light chain CDR3 of SEQ ID NO: 36 (QQYNSYSMYT).

21. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37, and a VL chain comprising the amino acid sequence of SEQ ID NO: 38; (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39, and a VL chain comprising the amino acid sequence of SEQ ID NO: 40; (c) a VH chain comprising the amino acid sequence of SEQ ID NO: 41, and a VL chain comprising the amino acid sequence of SEQ ID NO: 42; (d) a VH chain comprising the amino acid sequence of SEQ ID NO: 43, and a VL chain comprising the amino acid sequence of SEQ ID NO: 44; (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45, and a VL chain comprising the amino acid sequence of SEQ ID NO: 46; and (f) a VH chain comprising the amino acid sequence of SEQ ID NO:47, and a VL chain comprising the amino acid sequence of SEQ ID NO:

48.

22. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a heavy chain amino acid sequence of SEQ ID NO: 49, and a light chain amino acid sequence of SEQ ID NO: 50; (b) a heavy chain amino acid sequence of SEQ ID NO:51, and a light chain amino acid sequence of SEQ ID NO:52; (c) a heavy chain amino acid sequence of SEQ ID NO:53, and a light chain amino acid sequence of SEQ ID NO:54; (d) a heavy chain amino acid sequence of SEQ ID NO:55, and a light chain amino acid sequence of SEQ ID NO:56; (e) a heavy chain amino acid sequence of SEQ ID NO:57, and a light chain amino acid sequence of SEQ ID NO:58; and (f) a heavy chain amino acid sequence of SEQ ID NO:59, and a light chain amino acid sequence of SEQ ID NO:

60.

23. the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, a heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, a heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, a light chain CDR2 (KASGLES) of SEQ ID NO: 5, and a light chain CDR3 (QQYNSYSSQLT) of SEQ ID NO: 6; The ADC payload is a topoisomerase I inhibitor. 【Chemistry 4】 and 2. The method of claim 1 , wherein the ADC has a drug-to-antibody ratio (DAR) of 8.

24. The method of claim 23, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO:37 and a VL chain comprising the amino acid sequence of SEQ ID NO:

38.

25. 1. A method of generating a predictive efficacy score to predict a cancer patient's response to an antibody drug conjugate (ADC), comprising an ADC payload and an ADC antibody or antigen-binding fragment thereof that targets a protein on a cancer cell, the protein being folate receptor alpha (FRα), the method comprising: immunohistochemically staining a tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample; acquiring a digital image of the tissue sample; detecting cancer cells in the digital image; and determining for each cancer cell the average optical density of staining with the dye in the membrane of said cancer cell; determining a median optical density of all cancer cells in said digital image; generating a predicted efficacy score for the tissue sample based on the median optical density, the predicted efficacy score being positive if the median optical density is equal to or greater than an optical density threshold and being negative if the median optical density is less than the optical density threshold, the optical density threshold correlating with the response of a cohort of training patients treated with the ADC; and if the predicted efficacy score is positive, recommending a therapy involving the ADC to the cancer patient.

26. 26. The method of claim 25, further comprising administering the therapy involving the ADC to the cancer patient if the predicted efficacy score is positive.

27. 26. The method of claim 25, wherein said detecting cancer cells involves detecting, for each cancer cell, pixels belonging to the membrane using a cell centroid determined for each cancer cell.

28. 26. The method of claim 25, wherein the staining intensity of each membrane is calculated based on the average optical density of the brown diaminobenzidine (DAB) signal in the pixels of the membrane.

29. 26. The method of claim 25, wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

30. The method of claim 25, wherein the dye is 3,3'-diaminobenzidine (DAB).

31. 26. The method of claim 25, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer, and malignant pleural mesothelioma.

32. 26. The method of claim 25, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

33. 33. The method of claim 32, wherein the NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous and adenocarcinoma NSCLC.

34. 26. The method of claim 25, wherein the ADC payload is a cytotoxin.

35. 35. The method of claim 34, wherein the cytotoxin is a topoisomerase I inhibitor.

36. The topoisomerase I inhibitor has the following formula: 【Chemistry 5】 and salts and solvates thereof, wherein R Ｌ represents a linker connected to the ADC antibody or antigen-binding fragment thereof.

37. The ADC is an anti-FRα antibody, or antigen-binding fragment thereof, conjugated to a topoisomerase I inhibitor, the topoisomerase I inhibitor being 【Chemistry 6】 26. The method of claim 25, wherein the

38. The ADC is an anti-FRα antibody, or antigen-binding fragment thereof, conjugated to a topoisomerase I inhibitor, the topoisomerase I inhibitor being 【Chemistry 7】 26. The method of claim 25, wherein:

39. 26. The method of claim 25, wherein the ADC has a drug-to-antibody ratio (DAR) within a range selected from the group consisting of 1-20, 1-10, 2-10, 2-8, 2-6, and 4-10.

40. 26. The method of claim 25, wherein the ADC has a drug-to-antibody ratio (DAR) selected from the group consisting of 4 and 8.

41. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, a heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, a heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, a light chain CDR2 (KASGLES) of SEQ ID NO: 5, and a light chain CDR3 (QQYNSYSSQLT) of SEQ ID NO: 6; (b) a heavy chain CDR1 (SYAMS) of SEQ ID NO: 7, a heavy chain CDR2 (SISSGRSYIYYADSVKG) of SEQ ID NO: 8, a heavy chain CDR3 (EMQQLALDY) of SEQ ID NO: 9, a light chain CDR1 (RASQGISNFLA) of SEQ ID NO: 10, a light chain CDR2 (AASSLQS) of SEQ ID NO: 11, and a light chain CDR3 (QQYNSYPFT) of SEQ ID NO: 12; (c) a heavy chain CDR1 (SNSAAWN) of SEQ ID NO: 13, a heavy chain CDR2 (RTYYRSNWYNDYTLSVKS) of SEQ ID NO: 14, a heavy chain CDR3 (GVGRFDS) of SEQ ID NO: 15, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 16, a light chain CDR2 (KASSLES) of SEQ ID NO: 17, and a light chain CDR3 (QEYKTYSIFT) of SEQ ID NO: 18; (d) a heavy chain CDR1 (SYNMN) of SEQ ID NO: 19, a heavy chain CDR2 (SISSSGSSYIYYADSMKG) of SEQ ID NO: 20, a heavy chain CDR3 (GMTTLTFDY) of SEQ ID NO: 21, a light chain CDR1 (RASQGISTFLA) of SEQ ID NO: 22, a light chain CDR2 (AASSLQS) of SEQ ID NO: 23, and a light chain CDR3 (QQYISYPLT) of SEQ ID NO: 24; (e) a heavy chain CDR1 of SEQ ID NO: 25 (SYSMN), a heavy chain CDR2 of SEQ ID NO: 26 (SISSRSSYVYYADSVKG), a heavy chain CDR3 of SEQ ID NO: 27 (GMTTLTFDY), a light chain CDR1 of SEQ ID NO: 28 (RASQGISSFLA), a light chain CDR2 of SEQ ID NO: 29 (AASSLQS), and a light chain CDR3 of SEQ ID NO: 30 (QQYNSYPLT), and (f) the method of claim 25, comprising a plurality of amino acid sequences selected from heavy chain CDR1 of SEQ ID NO: 31 (SDSATWN), heavy chain CDR2 of SEQ ID NO: 32 (RTYYRSKWYSDYAVSVKS), heavy chain CDR3 of SEQ ID NO: 33 (GGAPFDY), light chain CDR1 of SEQ ID NO: 34 (RASQSISSWLA), light chain CDR2 of SEQ ID NO: 35 (KASSLES), and light chain CDR3 of SEQ ID NO: 36 (QQYNSYSMYT).

42. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37, and a VL chain comprising the amino acid sequence of SEQ ID NO: 38; (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39, and a VL chain comprising the amino acid sequence of SEQ ID NO: 40; (c) a VH chain comprising the amino acid sequence of SEQ ID NO: 41, and a VL chain comprising the amino acid sequence of SEQ ID NO: 42; (d) a VH chain comprising the amino acid sequence of SEQ ID NO: 43, and a VL chain comprising the amino acid sequence of SEQ ID NO: 44; (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45, and a VL chain comprising the amino acid sequence of SEQ ID NO: 46; and (f) a VH chain comprising the amino acid sequence of SEQ ID NO: 47, and a VL chain comprising the amino acid sequence of SEQ ID NO:

48. The method of claim 25, comprising a plurality of amino acid sequences selected from the group consisting of:

43. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a heavy chain amino acid sequence of SEQ ID NO: 49, and a light chain amino acid sequence of SEQ ID NO: 50; (b) a heavy chain amino acid sequence of SEQ ID NO:51, and a light chain amino acid sequence of SEQ ID NO:52; (c) a heavy chain amino acid sequence of SEQ ID NO:53, and a light chain amino acid sequence of SEQ ID NO:54; (d) a heavy chain amino acid sequence of SEQ ID NO:55, and a light chain amino acid sequence of SEQ ID NO:56; (e) a heavy chain amino acid sequence of SEQ ID NO:57, and a light chain amino acid sequence of SEQ ID NO:58; and (f) a heavy chain amino acid sequence of SEQ ID NO:59, and a light chain amino acid sequence of SEQ ID NO:

60.

44. the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, a heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, a heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, a light chain CDR2 (KASGLES) of SEQ ID NO: 5, and a light chain CDR3 (QQYNSYSSQLT) of SEQ ID NO: 6; The ADC payload is a topoisomerase I inhibitor. 【Chemistry 8】 and 26. The method of claim 25, wherein the ADC has a drug-to-antibody ratio (DAR) of 8.

45. The method of claim 44, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO:37 and a VL chain comprising the amino acid sequence of SEQ ID NO:

38.

46. 1. A method of generating a predictive efficacy score to predict a cancer patient's response to an antibody drug conjugate (ADC), comprising an ADC payload and an ADC antibody that targets a protein on a cancer cell, the protein being folate receptor alpha (FRα), the method comprising: immunohistochemically staining a tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample; acquiring a digital image of the tissue sample; detecting cancer cells in the digital image; and determining for each cancer cell the optical density of staining of the dye in the membrane of the cancer cell; identifying each cancer cell as either optical density positive if the optical density of the cancer cell is equal to or greater than an optical density threshold, or optical density negative if the mean optical density of the cancer cells is less than the optical density threshold; generating a predicted efficacy score for the tissue sample based on the percentage of cancer cells in the digital image that are optical density positive, the predicted efficacy score being positive if the percentage of the cancer cells that are optical density positive is equal to or greater than a percentage threshold and being negative if the percentage of the cancer cells that are optical density positive is less than the percentage threshold, the optical density threshold and the percentage threshold correlating with the response of a cohort of training patients treated with the ADC; and if the predicted efficacy score is positive, recommending a therapy involving the ADC to the cancer patient.

47. 47. The method of claim 46, further comprising administering the therapy involving the ADC to the cancer patient if the predicted efficacy score is positive.

48. 47. The method of claim 46, wherein the optical density threshold is in the range of 10 to 15 on a scale having a maximum optical density of 220.

49. 47. The method of claim 46, wherein the percentage threshold is in the range of 80% to 90%.

50. 47. The method of claim 46, wherein said detecting cancer cells involves detecting, for each cancer cell, pixels belonging to the membrane using a cell centroid determined for each cancer cell.

51. 47. The method of claim 46, wherein the staining intensity of each membrane is calculated based on the average optical density of the brown diaminobenzidine (DAB) signal in the pixels of the membrane.

52. 47. The method of claim 46, wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

53. 47. The method of claim 46, wherein the dye is 3,3'-diaminobenzidine (DAB).

54. 47. The method of claim 46, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer, and malignant pleural mesothelioma.

55. 47. The method of claim 46, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

56. 56. The method of claim 55, wherein the NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous and adenocarcinoma NSCLC.

57. 47. The method of claim 46, wherein the ADC payload is a cytotoxin.

58. 58. The method of claim 57, wherein the cytotoxin is a topoisomerase I inhibitor.

59. The topoisomerase I inhibitor has the following formula: 【Chemistry 9】 and salts and solvates thereof, wherein R Ｌ represents a linker connected to the ADC antibody or antigen-binding fragment thereof.

60. The ADC is an anti-FRα antibody, or antigen-binding fragment thereof, conjugated to a topoisomerase I inhibitor, the topoisomerase I inhibitor being 【Chemistry 10】 47. The method of claim 46, wherein said compound is selected from the group consisting of

61. The ADC is an anti-FRα antibody, or antigen-binding fragment thereof, conjugated to a topoisomerase I inhibitor, the topoisomerase I inhibitor being 【Chemistry 11】 47. The method of claim 46, wherein:

62. 47. The method of claim 46, wherein the ADC has a drug-to-antibody ratio (DAR) within a range selected from the group consisting of 1-20, 1-10, 2-10, 2-8, 2-6, and 4-10.

63. 47. The method of claim 46, wherein the ADC has a drug-to-antibody ratio (DAR) selected from the group consisting of 4 and 8.

64. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, a heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, a heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, a light chain CDR2 (KASGLES) of SEQ ID NO: 5, and a light chain CDR3 (QQYNSYSSQLT) of SEQ ID NO: 6; (b) a heavy chain CDR1 (SYAMS) of SEQ ID NO: 7, a heavy chain CDR2 (SISSGRSYIYYADSVKG) of SEQ ID NO: 8, a heavy chain CDR3 (EMQQLALDY) of SEQ ID NO: 9, a light chain CDR1 (RASQGISNFLA) of SEQ ID NO: 10, a light chain CDR2 (AASSLQS) of SEQ ID NO: 11, and a light chain CDR3 (QQYNSYPFT) of SEQ ID NO: 12; (c) a heavy chain CDR1 (SNSAAWN) of SEQ ID NO: 13, a heavy chain CDR2 (RTYYRSNWYNDYTLSVKS) of SEQ ID NO: 14, a heavy chain CDR3 (GVGRFDS) of SEQ ID NO: 15, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 16, a light chain CDR2 (KASSLES) of SEQ ID NO: 17, and a light chain CDR3 (QEYKTYSIFT) of SEQ ID NO: 18; (d) a heavy chain CDR1 (SYNMN) of SEQ ID NO: 19, a heavy chain CDR2 (SISSSGSSYIYYADSMKG) of SEQ ID NO: 20, a heavy chain CDR3 (GMTTLTFDY) of SEQ ID NO: 21, a light chain CDR1 (RASQGISTFLA) of SEQ ID NO: 22, a light chain CDR2 (AASSLQS) of SEQ ID NO: 23, and a light chain CDR3 (QQYISYPLT) of SEQ ID NO: 24; (e) a heavy chain CDR1 of SEQ ID NO: 25 (SYSMN), a heavy chain CDR2 of SEQ ID NO: 26 (SISSRSSYVYYADSVKG), a heavy chain CDR3 of SEQ ID NO: 27 (GMTTLTFDY), a light chain CDR1 of SEQ ID NO: 28 (RASQGISSFLA), a light chain CDR2 of SEQ ID NO: 29 (AASSLQS), and a light chain CDR3 of SEQ ID NO: 30 (QQYNSYPLT), and (f) the method of claim 46, comprising a plurality of amino acid sequences selected from heavy chain CDR1 of SEQ ID NO: 31 (SDSATWN), heavy chain CDR2 of SEQ ID NO: 32 (RTYYRSKWYSDYAVSVKS), heavy chain CDR3 of SEQ ID NO: 33 (GGAPFDY), light chain CDR1 of SEQ ID NO: 34 (RASQSISSWLA), light chain CDR2 of SEQ ID NO: 35 (KASSLES), and light chain CDR3 of SEQ ID NO: 36 (QQYNSYSMYT).

65. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37, and a VL chain comprising the amino acid sequence of SEQ ID NO: 38; (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39, and a VL chain comprising the amino acid sequence of SEQ ID NO: 40; (c) a VH chain comprising the amino acid sequence of SEQ ID NO: 41, and a VL chain comprising the amino acid sequence of SEQ ID NO: 42; (d) a VH chain comprising the amino acid sequence of SEQ ID NO: 43, and a VL chain comprising the amino acid sequence of SEQ ID NO: 44; (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45, and a VL chain comprising the amino acid sequence of SEQ ID NO: 46; and (f) a VH chain comprising the amino acid sequence of SEQ ID NO: 47, and a VL chain comprising the amino acid sequence of SEQ ID NO:

48. The method of claim 46, comprising a plurality of amino acid sequences selected from the group consisting of:

66. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a heavy chain amino acid sequence of SEQ ID NO: 49, and a light chain amino acid sequence of SEQ ID NO: 50; (b) a heavy chain amino acid sequence of SEQ ID NO:51, and a light chain amino acid sequence of SEQ ID NO:52; (c) a heavy chain amino acid sequence of SEQ ID NO:53, and a light chain amino acid sequence of SEQ ID NO:54; (d) a heavy chain amino acid sequence of SEQ ID NO:55, and a light chain amino acid sequence of SEQ ID NO:56; (e) a heavy chain amino acid sequence of SEQ ID NO:57, and a light chain amino acid sequence of SEQ ID NO:58; and (f) a heavy chain amino acid sequence of SEQ ID NO:59, and a light chain amino acid sequence of SEQ ID NO:

60.

67. the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, a heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, a heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, a light chain CDR2 (KASGLES) of SEQ ID NO: 5, and a light chain CDR3 (QQYNSYSSQLT) of SEQ ID NO: 6; The ADC payload is a topoisomerase I inhibitor. 【Chemistry 12】 and 47. The method of claim 46, wherein the ADC has a drug-to-antibody ratio (DAR) of 8.

68. The method of claim 67, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO:37 and a VL chain comprising the amino acid sequence of SEQ ID NO:

38.

69. 1. A method of identifying a cancer patient for treatment with an antibody drug conjugate (ADC), comprising an ADC payload and an ADC antibody that targets a protein on a cancer cell, the protein being folate receptor alpha (FRα), the method comprising: immunohistochemically staining a tissue sample from the cancer patient using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample; acquiring a digital image of the tissue sample; detecting cancer cells in the digital image; and determining for each cancer cell the average optical density of staining with the dye in the membrane of said cancer cell; identifying each cancer cell as either optical density positive if the mean optical density of the cancer cells is equal to or greater than an optical density threshold, or optical density negative if the mean optical density of the cancer cells is less than the optical density threshold; generating a proximity score for the tissue sample equal to the percentage of cancer cells in the digital image that are either optical density positive or optical density negative, but within a predetermined distance from optical density positive cancer cells; and if the proximity score exceeds a predetermined percentage threshold, identifying the cancer patient as one who would likely benefit from administration of the ADC.

70. 70. The method of claim 69, further comprising recommending therapy involving the ADC to the cancer patient if the proximity score exceeds the predetermined percentage threshold.

71. 70. The method of claim 69, further comprising administering to the cancer patient the therapy involving the ADC if the proximity score exceeds the predetermined percentage threshold.

72. 70. The method of claim 69, wherein the optical density threshold is in the range of 10 to 15 on a scale having a maximum optical density of 220.

73. 70. The method of claim 69, wherein the percentage threshold is in the range of 95% to 100%.

74. 70. The method of claim 69, wherein said detecting cancer cells involves detecting, for each cancer cell, pixels belonging to the membrane using a cell centroid determined for each cancer cell.

75. 70. The method of claim 69, wherein the staining intensity of each membrane is calculated based on the average optical density of the brown diaminobenzidine (DAB) signal in the pixels of the membrane.

76. 70. The method of claim 69, wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

77. 70. The method of claim 69, wherein the dye is 3,3'-diaminobenzidine (DAB).

78. 70. The method of claim 69, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer, and malignant pleural mesothelioma.

79. 70. The method of claim 69, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

80. 80. The method of claim 79, wherein the NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous and adenocarcinoma NSCLC.

81. 70. The method of claim 69, wherein the ADC payload is a cytotoxin.

82. 82. The method of claim 81, wherein the cytotoxin is a topoisomerase I inhibitor.

83. The topoisomerase I inhibitor has the following formula: 【Chemistry 13】 and salts and solvates thereof, wherein R Ｌ represents a linker connected to the ADC antibody or antigen-binding fragment thereof.

84. The ADC is an anti-FRα antibody, or antigen-binding fragment thereof, conjugated to a topoisomerase I inhibitor, the topoisomerase I inhibitor being 【Chemistry 14】 70. The method of claim 69, wherein said compound is selected from

85. The ADC is an anti-FRα antibody, or antigen-binding fragment thereof, conjugated to a topoisomerase I inhibitor, the topoisomerase I inhibitor being 【Chemistry 15】 70. The method of claim 69, wherein:

86. 70. The method of claim 69, wherein the ADC has a drug-to-antibody ratio (DAR) within a range selected from the group consisting of 1-20, 1-10, 2-10, 2-8, 2-6, and 4-10.

87. 70. The method of claim 69, wherein the ADC has a drug-to-antibody ratio (DAR) selected from the group consisting of 4 and 8.

88. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, a heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, a heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, a light chain CDR2 (KASGLES) of SEQ ID NO: 5, and a light chain CDR3 (QQYNSYSSQLT) of SEQ ID NO: 6; (b) a heavy chain CDR1 (SYAMS) of SEQ ID NO: 7, a heavy chain CDR2 (SISSGRSYIYYADSVKG) of SEQ ID NO: 8, a heavy chain CDR3 (EMQQLALDY) of SEQ ID NO: 9, a light chain CDR1 (RASQGISNFLA) of SEQ ID NO: 10, a light chain CDR2 (AASSLQS) of SEQ ID NO: 11, and a light chain CDR3 (QQYNSYPFT) of SEQ ID NO: 12; (c) a heavy chain CDR1 (SNSAAWN) of SEQ ID NO: 13, a heavy chain CDR2 (RTYYRSNWYNDYTLSVKS) of SEQ ID NO: 14, a heavy chain CDR3 (GVGRFDS) of SEQ ID NO: 15, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 16, a light chain CDR2 (KASSLES) of SEQ ID NO: 17, and a light chain CDR3 (QEYKTYSIFT) of SEQ ID NO: 18; (d) a heavy chain CDR1 (SYNMN) of SEQ ID NO: 19, a heavy chain CDR2 (SISSSGSSYIYYADSMKG) of SEQ ID NO: 20, a heavy chain CDR3 (GMTTLTFDY) of SEQ ID NO: 21, a light chain CDR1 (RASQGISTFLA) of SEQ ID NO: 22, a light chain CDR2 (AASSLQS) of SEQ ID NO: 23, and a light chain CDR3 (QQYISYPLT) of SEQ ID NO: 24; (e) a heavy chain CDR1 of SEQ ID NO: 25 (SYSMN), a heavy chain CDR2 of SEQ ID NO: 26 (SISSRSSYVYYADSVKG), a heavy chain CDR3 of SEQ ID NO: 27 (GMTTLTFDY), a light chain CDR1 of SEQ ID NO: 28 (RASQGISSFLA), a light chain CDR2 of SEQ ID NO: 29 (AASSLQS), and a light chain CDR3 of SEQ ID NO: 30 (QQYNSYPLT), and (f) the method of claim 69, comprising a plurality of amino acid sequences selected from heavy chain CDR1 of SEQ ID NO: 31 (SDSATWN), heavy chain CDR2 of SEQ ID NO: 32 (RTYYRSKWYSDYAVSVKS), heavy chain CDR3 of SEQ ID NO: 33 (GGAPFDY), light chain CDR1 of SEQ ID NO: 34 (RASQSISSWLA), light chain CDR2 of SEQ ID NO: 35 (KASSLES), and light chain CDR3 of SEQ ID NO: 36 (QQYNSYSMYT).

89. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37, and a VL chain comprising the amino acid sequence of SEQ ID NO: 38; (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39, and a VL chain comprising the amino acid sequence of SEQ ID NO: 40; (c) a VH chain comprising the amino acid sequence of SEQ ID NO: 41, and a VL chain comprising the amino acid sequence of SEQ ID NO: 42; (d) a VH chain comprising the amino acid sequence of SEQ ID NO: 43, and a VL chain comprising the amino acid sequence of SEQ ID NO: 44; (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45, and a VL chain comprising the amino acid sequence of SEQ ID NO: 46; and (f) a VH chain comprising the amino acid sequence of SEQ ID NO: 47, and a VL chain comprising the amino acid sequence of SEQ ID NO:

48. The method of claim 69, comprising a plurality of amino acid sequences selected from the group consisting of:

90. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a heavy chain amino acid sequence of SEQ ID NO: 49, and a light chain amino acid sequence of SEQ ID NO: 50; (b) a heavy chain amino acid sequence of SEQ ID NO:51, and a light chain amino acid sequence of SEQ ID NO:52; (c) a heavy chain amino acid sequence of SEQ ID NO:53, and a light chain amino acid sequence of SEQ ID NO:54; (d) a heavy chain amino acid sequence of SEQ ID NO:55, and a light chain amino acid sequence of SEQ ID NO:56; (e) a heavy chain amino acid sequence of SEQ ID NO:57, and a light chain amino acid sequence of SEQ ID NO:58; and (f) a heavy chain amino acid sequence of SEQ ID NO:59, and a light chain amino acid sequence of SEQ ID NO:

60.

91. the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, a heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, a heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, a light chain CDR2 (KASGLES) of SEQ ID NO: 5, and a light chain CDR3 (QQYNSYSSQLT) of SEQ ID NO: 6; The ADC payload is a topoisomerase I inhibitor. 【Chemistry 16】 and 70. The method of claim 69, wherein the ADC has a drug-to-antibody ratio (DAR) of 8.

92. The method of claim 91, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO:37 and a VL chain comprising the amino acid sequence of SEQ ID NO:

38.

93. 1. A method of generating a predictive efficacy score to predict a cancer patient's response to an antibody drug conjugate (ADC), comprising an ADC payload and an ADC antibody that targets a protein on a cancer cell, the protein being folate receptor alpha (FRα), the method comprising: immunohistochemically staining a tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample; acquiring a digital image of the tissue sample; detecting cancer cells in the digital image; and determining for each cancer cell the average optical density of staining with the dye in the membrane of said cancer cell; determining a median optical density of all cancer cells in said digital image; determining a median absolute deviation of the optical density of the cancer cells from the median of the optical density of all cancer cells in the digital image; generating a predicted efficacy score for the tissue sample based on the median absolute deviation, the predicted efficacy score being positive if the median absolute deviation is equal to or greater than a deviation threshold and being negative if the median absolute deviation is less than the deviation threshold, the deviation threshold correlating with a response of a cohort of training patients treated with the ADC; and if the predicted efficacy score is positive, recommending a therapy involving the ADC to the cancer patient.

94. 94. The method of claim 93, further comprising administering said therapy involving said ADC to said cancer patient if said predicted efficacy score is positive.

95. 94. The method of claim 93, wherein the deviation threshold is in the range of 10 to 15 on an optical density scale having a maximum value of 220.

96. 94. The method of claim 93, wherein said detecting cancer cells involves detecting, for each cancer cell, pixels belonging to the membrane using a cell centroid determined for each cancer cell.

97. 94. The method of claim 93, wherein the staining intensity of each membrane is calculated based on the average optical density of the brown diaminobenzidine (DAB) signal in the pixels of the membrane.

98. 94. The method of claim 93, wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

99. 94. The method of claim 93, wherein the dye is 3,3'-diaminobenzidine (DAB).

100. 94. The method of claim 93, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer, and malignant pleural mesothelioma.

101. 94. The method of claim 93, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

102. 102. The method of claim 101, wherein the NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous NSCLC and adenocarcinoma NSCLC.

103. 94. The method of claim 93, wherein the ADC payload is a cytotoxin.

104. 104. The method of claim 103, wherein the cytotoxin is a topoisomerase I inhibitor.

105. The topoisomerase I inhibitor has the following formula: 【Chemistry 17】 and salts and solvates thereof, wherein R Ｌ represents a linker connected to the ADC antibody or antigen-binding fragment thereof.

106. The ADC is an anti-FRα antibody, or antigen-binding fragment thereof, conjugated to a topoisomerase I inhibitor, the topoisomerase I inhibitor being 【Chemistry 18】 94. The method of claim 93, wherein said

107. The ADC is an anti-FRα antibody, or antigen-binding fragment thereof, conjugated to a topoisomerase I inhibitor, the topoisomerase I inhibitor being 【Chemistry 19】 94. The method of claim 93, wherein:

108. 94. The method of claim 93, wherein the ADC has a drug-to-antibody ratio (DAR) within a range selected from the group consisting of 1-20, 1-10, 2-10, 2-8, 2-6, and 4-10.

109. 94. The method of claim 93, wherein the ADC has a drug-to-antibody ratio (DAR) selected from the group consisting of 4 and 8.

110. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, a heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, a heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, a light chain CDR2 (KASGLES) of SEQ ID NO: 5, and a light chain CDR3 (QQYNSYSSQLT) of SEQ ID NO: 6; (b) a heavy chain CDR1 (SYAMS) of SEQ ID NO: 7, a heavy chain CDR2 (SISSGRSYIYYADSVKG) of SEQ ID NO: 8, a heavy chain CDR3 (EMQQLALDY) of SEQ ID NO: 9, a light chain CDR1 (RASQGISNFLA) of SEQ ID NO: 10, a light chain CDR2 (AASSLQS) of SEQ ID NO: 11, and a light chain CDR3 (QQYNSYPFT) of SEQ ID NO: 12; (c) a heavy chain CDR1 (SNSAAWN) of SEQ ID NO: 13, a heavy chain CDR2 (RTYYRSNWYNDYTLSVKS) of SEQ ID NO: 14, a heavy chain CDR3 (GVGRFDS) of SEQ ID NO: 15, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 16, a light chain CDR2 (KASSLES) of SEQ ID NO: 17, and a light chain CDR3 (QEYKTYSIFT) of SEQ ID NO: 18; (d) a heavy chain CDR1 (SYNMN) of SEQ ID NO: 19, a heavy chain CDR2 (SISSSGSSYIYYADSMKG) of SEQ ID NO: 20, a heavy chain CDR3 (GMTTLTFDY) of SEQ ID NO: 21, a light chain CDR1 (RASQGISTFLA) of SEQ ID NO: 22, a light chain CDR2 (AASSLQS) of SEQ ID NO: 23, and a light chain CDR3 (QQYISYPLT) of SEQ ID NO: 24; (e) a heavy chain CDR1 of SEQ ID NO: 25 (SYSMN), a heavy chain CDR2 of SEQ ID NO: 26 (SISSRSSYVYYADSVKG), a heavy chain CDR3 of SEQ ID NO: 27 (GMTTLTFDY), a light chain CDR1 of SEQ ID NO: 28 (RASQGISSFLA), a light chain CDR2 of SEQ ID NO: 29 (AASSLQS), and a light chain CDR3 of SEQ ID NO: 30 (QQYNSYPLT), and (f) the method of claim 93, comprising a plurality of amino acid sequences selected from heavy chain CDR1 of SEQ ID NO: 31 (SDSATWN), heavy chain CDR2 of SEQ ID NO: 32 (RTYYRSKWYSDYAVSVKS), heavy chain CDR3 of SEQ ID NO: 33 (GGAPFDY), light chain CDR1 of SEQ ID NO: 34 (RASQSISSWLA), light chain CDR2 of SEQ ID NO: 35 (KASSLES), and light chain CDR3 of SEQ ID NO: 36 (QQYNSYSMYT).

111. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37, and a VL chain comprising the amino acid sequence of SEQ ID NO: 38; (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39, and a VL chain comprising the amino acid sequence of SEQ ID NO: 40; (c) a VH chain comprising the amino acid sequence of SEQ ID NO: 41, and a VL chain comprising the amino acid sequence of SEQ ID NO: 42; (d) a VH chain comprising the amino acid sequence of SEQ ID NO: 43, and a VL chain comprising the amino acid sequence of SEQ ID NO: 44; (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45, and a VL chain comprising the amino acid sequence of SEQ ID NO: 46; and (f) a VH chain comprising the amino acid sequence of SEQ ID NO: 47, and a VL chain comprising the amino acid sequence of SEQ ID NO:

48. The method of claim 93, comprising a plurality of amino acid sequences selected from the group consisting of:

112. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a heavy chain amino acid sequence of SEQ ID NO: 49, and a light chain amino acid sequence of SEQ ID NO: 50; (b) a heavy chain amino acid sequence of SEQ ID NO:51, and a light chain amino acid sequence of SEQ ID NO:52; (c) a heavy chain amino acid sequence of SEQ ID NO:53, and a light chain amino acid sequence of SEQ ID NO:54; (d) a heavy chain amino acid sequence of SEQ ID NO:55, and a light chain amino acid sequence of SEQ ID NO:56; (e) a heavy chain amino acid sequence of SEQ ID NO:57, and a light chain amino acid sequence of SEQ ID NO:58; and (f) a heavy chain amino acid sequence of SEQ ID NO:59, and a light chain amino acid sequence of SEQ ID NO:

60.

113. the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, a heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, a heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, a light chain CDR2 (KASGLES) of SEQ ID NO: 5, and a light chain CDR3 (QQYNSYSSQLT) of SEQ ID NO: 6; The ADC payload is a topoisomerase I inhibitor. 【Chemistry 20】 and 94. The method of claim 93, wherein the ADC has a drug-to-antibody ratio (DAR) of 8.

114. The method of claim 113, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO: 37 and a VL chain comprising the amino acid sequence of SEQ ID NO:

38.

115. 1. A method of generating a predictive efficacy score to predict a cancer patient's response to an antibody drug conjugate (ADC), comprising an ADC payload and an ADC antibody that targets a protein on a cancer cell, the protein being folate receptor alpha (FRα), the method comprising: immunohistochemically staining a tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample; acquiring a digital image of the tissue sample; detecting cancer cells in the digital image; and determining for each cancer cell the average optical density of staining with the dye in the membrane of said cancer cell; determining for each cancer cell the average optical density of staining with the dye in the cytoplasm of said cancer cell; determining the difference between the average optical density of the membrane staining and the average optical density of the cytoplasmic staining for each cancer cell in the digital image; identifying an 85% quantile of the difference between the average optical density of the membrane staining and the average optical density of the cytoplasmic staining among all cancer cells in the digital image; and recommending a therapy involving the ADC to the cancer patient if the 85% quantile of the difference exceeds a predetermined difference threshold.

116. 116. The method of claim 115, further comprising administering to the cancer patient the therapy involving the ADC if the 85% quantile of the difference exceeds the predetermined difference threshold.

117. 116. The method of claim 115, wherein the difference threshold is in the range of 21 to 25 on an optical density scale having a maximum value of 220.

118. 116. The method of claim 115, wherein the detection of cancer cells involves detecting, for each cancer cell, pixels belonging to the membrane and pixels belonging to the cytoplasm.

119. 116. The method of claim 115, wherein the staining intensity of each membrane is calculated based on the average optical density of the brown diaminobenzidine (DAB) signal in the pixels of the membrane.

120. The method of claim 115, wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

121. 116. The method of claim 115, wherein the dye is 3,3'-diaminobenzidine (DAB).

122. 116. The method of claim 115, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer, and malignant pleural mesothelioma.

123. 116. The method of claim 115, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

124. 124. The method of claim 123, wherein the NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous and adenocarcinoma NSCLC.

125. The method of claim 115, wherein the ADC payload is a cytotoxin.

126. 126. The method of claim 125, wherein the cytotoxin is a topoisomerase I inhibitor.

127. The topoisomerase I inhibitor has the following formula: 【Chemistry 21】 and salts and solvates thereof, wherein R Ｌ represents a linker connected to the ADC antibody or antigen-binding fragment thereof.

128. The ADC is an anti-FRα antibody, or antigen-binding fragment thereof, conjugated to a topoisomerase I inhibitor, the topoisomerase I inhibitor being 【Chemical 22】 The method of claim 115, wherein the

129. The ADC is an anti-FRα antibody, or antigen-binding fragment thereof, conjugated to a topoisomerase I inhibitor, the topoisomerase I inhibitor being 【Chemistry 23】 The method of claim 115,

130. 116. The method of claim 115, wherein the ADC has a drug-to-antibody ratio (DAR) within a range selected from the group consisting of 1-20, 1-10, 2-10, 2-8, 2-6, and 4-10.

131. 116. The method of claim 115, wherein the ADC has a drug-to-antibody ratio (DAR) selected from the group consisting of 4 and 8.

132. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, a heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, a heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, a light chain CDR2 (KASGLES) of SEQ ID NO: 5, and a light chain CDR3 (QQYNSYSSQLT) of SEQ ID NO: 6; (b) a heavy chain CDR1 (SYAMS) of SEQ ID NO: 7, a heavy chain CDR2 (SISSGRSYIYYADSVKG) of SEQ ID NO: 8, a heavy chain CDR3 (EMQQLALDY) of SEQ ID NO: 9, a light chain CDR1 (RASQGISNFLA) of SEQ ID NO: 10, a light chain CDR2 (AASSLQS) of SEQ ID NO: 11, and a light chain CDR3 (QQYNSYPFT) of SEQ ID NO: 12; (c) a heavy chain CDR1 (SNSAAWN) of SEQ ID NO: 13, a heavy chain CDR2 (RTYYRSNWYNDYTLSVKS) of SEQ ID NO: 14, a heavy chain CDR3 (GVGRFDS) of SEQ ID NO: 15, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 16, a light chain CDR2 (KASSLES) of SEQ ID NO: 17, and a light chain CDR3 (QEYKTYSIFT) of SEQ ID NO: 18; (d) a heavy chain CDR1 (SYNMN) of SEQ ID NO: 19, a heavy chain CDR2 (SISSSGSSYIYYADSMKG) of SEQ ID NO: 20, a heavy chain CDR3 (GMTTLTFDY) of SEQ ID NO: 21, a light chain CDR1 (RASQGISTFLA) of SEQ ID NO: 22, a light chain CDR2 (AASSLQS) of SEQ ID NO: 23, and a light chain CDR3 (QQYISYPLT) of SEQ ID NO: 24; (e) a heavy chain CDR1 of SEQ ID NO: 25 (SYSMN), a heavy chain CDR2 of SEQ ID NO: 26 (SISSRSSYVYYADSVKG), a heavy chain CDR3 of SEQ ID NO: 27 (GMTTLTFDY), a light chain CDR1 of SEQ ID NO: 28 (RASQGISSFLA), a light chain CDR2 of SEQ ID NO: 29 (AASSLQS), and a light chain CDR3 of SEQ ID NO: 30 (QQYNSYPLT), and (f) the method of claim 115, comprising a plurality of amino acid sequences selected from heavy chain CDR1 of SEQ ID NO: 31 (SDSATWN), heavy chain CDR2 of SEQ ID NO: 32 (RTYYRSKWYSDYAVSVKS), heavy chain CDR3 of SEQ ID NO: 33 (GGAPFDY), light chain CDR1 of SEQ ID NO: 34 (RASQSISSWLA), light chain CDR2 of SEQ ID NO: 35 (KASSLES), and light chain CDR3 of SEQ ID NO: 36 (QQYNSYSMYT).

133. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37, and a VL chain comprising the amino acid sequence of SEQ ID NO: 38; (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39, and a VL chain comprising the amino acid sequence of SEQ ID NO: 40; (c) a VH chain comprising the amino acid sequence of SEQ ID NO: 41, and a VL chain comprising the amino acid sequence of SEQ ID NO: 42; (d) a VH chain comprising the amino acid sequence of SEQ ID NO: 43, and a VL chain comprising the amino acid sequence of SEQ ID NO: 44; (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45, and a VL chain comprising the amino acid sequence of SEQ ID NO: 46; and (f) a VH chain comprising the amino acid sequence of SEQ ID NO: 47, and a VL chain comprising the amino acid sequence of SEQ ID NO:

48. The method of claim 115, comprising a plurality of amino acid sequences selected from the group consisting of:

134. The ADC antibody or antigen-binding fragment thereof is selected from the group consisting of: (a) a heavy chain amino acid sequence of SEQ ID NO: 49, and a light chain amino acid sequence of SEQ ID NO: 50; (b) a heavy chain amino acid sequence of SEQ ID NO:51, and a light chain amino acid sequence of SEQ ID NO:52; (c) a heavy chain amino acid sequence of SEQ ID NO:53, and a light chain amino acid sequence of SEQ ID NO:54; (d) a heavy chain amino acid sequence of SEQ ID NO:55, and a light chain amino acid sequence of SEQ ID NO:56; (e) a heavy chain amino acid sequence of SEQ ID NO:57, and a light chain amino acid sequence of SEQ ID NO:58; and (f) a heavy chain amino acid sequence of SEQ ID NO:59, and a light chain amino acid sequence of SEQ ID NO:

60.

135. the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 (SDSATWN) of SEQ ID NO: 1, a heavy chain CDR2 (RTYYRSKWYNDYAVSVKS) of SEQ ID NO: 2, a heavy chain CDR3 (GVGSFDY) of SEQ ID NO: 3, a light chain CDR1 (RASQSISSWLA) of SEQ ID NO: 4, a light chain CDR2 (KASGLES) of SEQ ID NO: 5, and a light chain CDR3 (QQYNSYSSQLT) of SEQ ID NO: 6; The ADC payload is a topoisomerase I inhibitor. 【Chemistry 24】 and The method of claim 115, wherein the ADC has a drug-to-antibody ratio (DAR) of 8.

136. The method of claim 135, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO: 37 and a VL chain comprising the amino acid sequence of SEQ ID NO: 38.

Images