WO2025137381A1 - Using extracellular vesicles to assess treatment efficacy of poly (adp-ribose) polymerase inhibitor therapies - Google Patents

Abstract

The present disclosure relates to methods of detecting and analyzing tumor resistance, e.g., over time, to PARPi drugs used to treat the caner.

Description

USING EXTRACELLULAR VESICLES TO ASSESS TREATMENT EFFICACY

OF POLY (ADP-RIBOSE) POLYMERASE INHIBITOR THERAPIES

CLAIM OF PRIORITY

This application claims the benefit of U.S. Provisional Patent Application Serial No. 63/611,947, filed on December 19, 2023. The entire contents of the foregoing are hereby incorporated by reference.

FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

This invention was made with Government support under Grant Nos. R21CA217662 and R01GM138778 awarded by the National Institutes of Health. The Government has certain rights in the invention.

FIELD OF THE INVENTION

The disclosure is in the field of assessing the efficacy of cancer treatment using extracellular vesicles.

BACKGROUND OF THE INVENTION

Ovarian cancer (OC) is one of the most lethal gynecologic cancers, with an estimates 21,410 cases and 13,770 deaths in the United States in 2021. The standard therapy for ovarian cancer has been the same for the past two decades: a combination treatment of platinum with paclitaxel (Damia G. et al., 2019, Cancers (Basel);l 1(1): 119. doi: 10.3390/cancersl 1010119). Recently, olaparib (e.g., Lynparza®, AstraZeneca) a poly (ADP-ribose) polymerase inhibitor (PARPi), has been added to the standard of care for OC patients both front-line and recunent settings. Olaparib is also indicated to treat human epidermal growth factor receptor 2 (HER2)-negative, high-risk breast cancer, gBRCAm metastatic pancreatic adenocarcinoma, and germline or somatic homologous recombination repair (HRR) gene-mutated metastatic castration-resistant prostate cancer (mCRPC). However, olaparib's initial therapeutic promise is compromised by the mounting issue of resistance, posing a threat to its sustained effectiveness. Studies suggest that 40-70% of patients with ovarian cancer may develop resistance during first-line treatment (Kim et al., PARP Inhibitors: Clinical Limitations and Recent Attempts to Overcome Them. Int J Mol Sci 23, 8412 (2022)), while approximately 20-30% might face resistance within a five-year duration during maintenance therapy (DiSilvestro et al., J Clin. Oncol., 41(3):609-617 (2023)).

The mechanisms underlying olaparib resistance are complex and remain incompletely understood. These mechanisms encompass enhanced DNA repair pathways like homologous recombination (HR), non-homologous end joining (NHEJ), alternative end joining (alt-EJ), secondary mutations in DNA repair genes, alterations in drug efflux, changes in drug target sites, and activation of alternative DNA damage response pathways (Giudice et al., Cancers (Basel), 10;14(6): 1420. doi: 10.3390/cancersl4061420 (2022)). Anticipating the activation of these mechanisms and foreseeing a potential lack of response or reduced treatment effectiveness poses significant challenges. Thus, the desperate unmet need is a simple, accurate measurement of olaparib efficacy and tumor resistance that will help determine whether keeping a patient on PARPi treatment or pivoting to another option will significantly improve patient care, minimizing ineffective treatments.

SUMMARY

To address this unmet need, we developed a liquid biopsy technology that analyzes circulating extracellular vesicles (EVs). EVs carry biomolecules originating from their parental cells. Thus, molecular analysis of tumor-derived EVs (tEVs) reflect the molecular status of originating tumor cells. Considering EVs are actively shed by tumor cells, these unique properties place EVs as prominent biomarkers for longitudinal treatment monitoring using the methods described herein. Molecular EV analysis through liquid biopsy provides unique opportunities to monitor temporal changes of tumors during therapy in a non-invasive manner. This also provides more frequent assessment and in-depth molecular profiling than currently available procedures based on imaging and biopsy.

The present disclosure provides methods of identifying treatment resistant cancers, either solid tumors or hematologic malignancies, in a patient. For example, the methods can be applied to patients being treated for a tumor, e.g., by PARP inhibitors. The PARP inhibitors can include, without limitation, Olaparib, Rucaparib, Niraparib and Talazoparib.

Accordingly, in one aspect, the present disclosure provides methods of monitoring a tumor’s resistance to a poly (ADP -ribose) polymerase inhibitor (PARPi) drug administered to a subject to treat the tumor, the method including (a) obtaining a sample from the subject at a first time point; (b) isolating and quantifying extracellular vesicles (EVs) from the sample based on detection of EV biomarkers on the EVs in the sample; and (c) determining a presence or level of a tumor drug-resistance biomarker of the EVs isolated at the first time point, wherein a presence or level of the tumor drug-resistance biomarker indicates a resistance by the tumor to the PARPi drug being administered to the patient.

In another aspect, the disclosure provides methods of monitoring a tumor’s resistance to a poly (ADP-ribose) polymerase inhibitor (PARPi) drug, the methods including (a) obtaining a sample containing cells from the tumor at a first time point; (b) isolating and quantifying extracellular vesicles (EVs) from the sample based on detection of EV biomarkers on the EVs in the sample; and (c) determining a presence or level of a tumor drug-resi stance biomarker of the EVs isolated at the first time point, wherein a presence or level of the tumor drug-resistance biomarker indicates a resistance by the tumor to the PARPi drug.

In embodiments of these methods, step (b) can include isolating tumor-derived EVs (tEVs) from the sample based on detection of tumor biomarkers on the tEVs from the sample. For example, in some embodiments, step (b) can include first isolating EVs from the sample, and then isolating tEVs from the EVs isolate from the sample, or the tEVs can be isolated directly from the sample.

In some embodiments, step (c) can include companng the presence or level of the tumor drug-resistance biomarker to a reference sample or reference level from a subject known to be healthy and/or cancer-free or to have cancer with a good response to the PARPi drug, wherein a presence of the tumor drug-resistance biomarker in the sample when there is no presence or low levels of the tumor drug-resistance biomarker in the reference, or when a level of the tumor drug-resistance biomarker that is higher in the sample than the reference level, indicates a resistance by the tumor to the PARPi drug.

In certain embodiments, the methods can further include (d) obtaining a sample, e.g., from the subject, at a second time point, which is after the first time point; (e) isolating EVs or tEVs from the sample based on detection of EV biomarkers on the EVs or tumor biomarkers on the tEVs at the second time point; (f) determining a presence or level of a tumor drug-resistance biomarker of the EVs or the tEVs isolated at the second time point; and (g) comparing a presence or level of the tumor drug-resistance biomarker at the first time point to a presence or level of the tumor drug-resistance biomarker at the second time point, wherein a difference in a presence or level between the first time point and the second time point indicates that the tumor’s resistance to the PARPi has changed over time.

In embodiments of these methods, a presence of the tumor drug-resistance biomarker at the second time point, but not the first time point indicates that the tumor has developed a resistance to the PARPi drug over time, and/or an increase in the level of the tumor drug-resi stance biomarker at the second time point compared to the level at the first time point indicates that the tumor’s resistance to the PARPi drug has increased over time.

In various implementations of the methods described herein, the first time point can be before treatment, and the second time point can be during treatment or after treatment. In other implementations, the first time point is after treatment has begun, and the second time point is during ongoing treatment.

In various embodiments, the sample can be or include a bodily fluid, such as a blood sample, or can include subject-derived organoids or cells or cells derived from a tumor. In certain implementations, the EVs or tEVs are captured on a nanosensor chip for processing.

In various implementations of the methods described herein, the EVs or tEVs are labeled, e.g., with a reporter group, such as a fluorescent reporter group, that are bound to antibodies that bind specifically to one or more tumor drug-resistance biomarkers. Antibodies that bind specifically to tumor drug-resistance biomarkers tend not to bind, or to bind at a much lower level, to other proteins or other antigens that are not tumor drugresistance biomarkers.

In the methods described herein, a change in one or more tumor drug-resistance biomarkers can be tracked over the course of a portion or all of the subject’s treatment.

In various embodiments, the PARP inhibitor can be selected from the group consisting of olapanb, rucaparib, niraparib, and/or talazoparib. In some embodiments, an additional anti-cancer drug, e.g., a chemotherapeutic drug, is added to the sample or is administered to the subject before a sample is obtained from the subject.

In another aspect, the methods described herein include monitoring a patient’s response to cancer drug administration, by steps including (a) isolating tEVs from a patient blood sample, and (b) determining whether the tEVs exhibit one or more biomarkers that indicate resistance to the cancer drug being administered to the patient. In these methods, the patient’s blood sample is taken before treatment, during treatment, and after treatment and the determination whether the tEVs exhibit one or more biomarkers indicating cancer drug resistance is done for each sample taken to investigate whether the isolated tEVs exhibit differential levels in the drug-resistance tumor cells. The tEVs are isolated by capture on a nanosensor chip and are labeled with antibodies that bind specifically to target drug-resistance markers. The tEVs can be isolated from in vitro models and from human samples and the samples compared and changes in the markers can be tracked over the course of the patient’s treatment.

The growing emphasis on personalized cancer therapy and precision medicine increases the need to establish reliable, highly specific clinical assays for real-time cancer treatment monitoring. The methods disclosed herein provide a unique opportunity to access molecular information of tumors through non-invasive, liquid biopsy assays more frequently during the course of the treatment, which is not currently possible. tEVs are suitable targets for treatment monitonng. TEVs carry biomolecules that reflect the molecular status of their originating tumors, and biomarkers in tEVs show how cells change upon drug treatment. Compared to rare circulating tumor cells (CTC), tEVs are much more abundant in circulation and address the heterogeneity of tumors from which CTC analysis suffers. Unlike soluble biomarkers (DNA, RNA, proteins), tEVs carry surface proteins representing their cellular origins. This allows us to differentially evaluate changes in the specific drug-resistant biomarkers from tEVs and non-tEVs. Better understanding EV proteome changes to PARPi response will revolutionize our understanding of resistance and capacity to detect it clinically. While focusing on PARPi, the success here will elevate EV assays for use as an omics-like tool and potential as a liquid biopsy for increased clinical success.

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 invention belongs. Although methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, suitable methods and materials are described below. All publications, patent applications, patents, and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control. In addition, the materials, methods, and examples are illustrative only and not intended to be limiting.

Other features and advantages of the invention will be apparent from the following detailed description, and from the claims. DESCRIPTION OF THE DRAWINGS

FIG. 1A is a schematic representation illustrating the process of drug-resistance cell establishment.

FIG. IB is a graph that shows dose-response curves of cell viability of the UWV 1.289 ovarian cell line (parental) and olaparib resistance cell line (REOL).

FIG. 1C is a graph that shows a dose-response curve of cell viability of the OVCA429 ovarian cell line (parental) and olaparib resistance cell line (REOL).

FIG. 2A is a bar graph showing comparative growth profiles between parental and olaparib-resistant cell subtypes in terms of absolute cell counts per well at designated time points with corresponding doubling time (in days) for each cell line outlined in the table.

FIGs. 2B-2C are a pair of bar graphs that show quantitative data illustrating cell cycle distribution. All values presented are mean± SD. *P < 0.05.

FIGs. 3A-3F are a series of graphs that show molecular profiling of DNA damage response markers on parental and resistant cell subtypes. DNA damage markers in nuclei after 24-hour olaparib treatment in parental (UWB 1.289 and OVCA429) and their corresponding olaparib-resistant subtypes (UWB1.289 REOL and OVCA429 REOL). (FIGs. 3A and 3B) yH2AX, (FIGs. 3C and 3D) RAD 1, and (FIG. 3E and 3F) PCNA. DAPI was utilized for nuclear segmentation, classifying cells into Gl, S, and G2 cell cycle phases. Normalization of marker intensity was performed relative to the DMSO control.

FIGs. 4A-4D are a series of graphs that show NTA analysis of cell line-derived EVs and the evaluation of the size repartition of the EVs secreted by the UWB1.289 (Parental) and UWB 1.289 REOL (FIG. 4A and 4B), as well as OVCA429 (Parental), and OVCA429 REOL (FIGs. 4C and 4D). Most EV populations were comprised between 50 and 180 nm, which is the size range commonly given to small EVs.

FIG. 5A is a Venn diagram that shows differentially expressed proteins in EV derived from parental and resistant cell subtype and the shared differential protein expression among each subtype of cell line-derived EVs.

FIG. 5B is a is a type of bar graph that shows Gene Ontology (GO) term enrichment analysis that compares EVs from UWB 1.289 parental and REOL cell subtypes. FIG. 5C is a type of bar graph that shows OVCA429 parental and REOL cell subtypes. Molecular Function (MF), Cellular Component (CC), and Biological Process (BP).

FIG. 6 is a heatmap of expression profiles that showsthe expression profiles of 20 commonly enriched differentially expressed proteins (DEPs) found in EVs isolated from both cellular models (UWB1.289 and OVCA429).

FIG. 7 is a schematic that shows that EVs from tumor cell lines labeled by TFP- AF555 are captured on a nanosensor chip substrate and immunolabeled by anti-INHBA antibodies labeled by AF647-conjugated secondary antibody and anti-AEBPl antibodies labeled by AF488 -conjugated secondary antibody. Co-localized signals of AF488 (AEBP1) or AF647 (INHBA) with AF555 (tEVs) define their presence and levels in the tEVs.

FIG. 7B is a bar graph that shows EV protein levels of candidate markers for olapanb resistance and colocalization percentage of inhibin, beta A, also known as INHBA, which is a protein that in humans is encoded by the INHBA gene. The data in the graph includes IGRb correction (a measure of protein marker abundance in EVs) for four different cell lines. Shown is an average with standard deviation and individual values (n = 3).

FIG. 7C is a bar graph that shows EV protein levels of candidate markers for olapanb resistance and colocalization percentage of AEBP1 with IGRb correction for four different cell lines. Shown is average with standard deviation and individual values (n = 3). AEBP1 is a protein that in humans is encoded by the AEBP1 gene. AEBP1 is a member of carboxypeptidase A protein family, which may function as a transcriptional repressor and play a role in adipogenesis and smooth muscle cell differentiation as well as in wound healing and abdominal wall development. Overexpression of this gene is associated with glioblastoma.

FIGs. 8A-8C are bar graphs showing INHBA-positive EV counts at three-time points before treatment, 4-6 weeks after treatments, and 3-6 months after treatments. PIPS are plasma samples from three patients who each received PARPi treatments and whose tumors progressed. P4-P6 are samples from three patients who received other drugs than PARPi and whose tumors progressed. The IHNBA-positive tEV counts increased only in samples from P1-P3. In particular, FIG. 8A shows EV counts (TFP- AF555), FIG. 8B shows the quantification of colocalization in EVs positive for the marker (INHBA) and EV channels, and FIG. 8C shows the percentage of EV colocalization (TFP-555+INHBA) in each sample.

FIG. 9 is a schematic that shows tEVs labeled by TFP-AF555 are captured on a nanosensor chip substrate by capture antibodies, a mixture of anti-EpCAM and anti- CD24. The captured EVs were immunolabeled by anti-INHBA antibodies labeled by AF647-conjugated secondary antibody and anti-AEBPl antibodies labeled by AF488- conjugated secondary antibody. Co-localized signals of AF488 (AEBP1) or AF647 (INHBA) with AF555 (tEVs) define their presence and levels in the tEVs.

FIG. 9B is a senes of bar graphs that show tEVs captured by anti-EpCAM and anti-CD24 immobilized on the nanosensor surface (EV Channel, left). tEVs were isolated from paired cell lines (UWB1.289 PA, OVCA429 PA, PEO-1) and their resistant models for olapanb (UWB1.289 PA REOL, OVCA429 PA REOL, PEO-1 REOL). The isolated tEVs were labeled by TFP-AF555. The captured EVs were then immunolabeled by AEBP1 followed by AF647-conjugated secondary antibody. The AF647 signals are shown in blue bars (Marker Channel, middle). EV counts with co-localized signals of AF555 (EVs) and AF647 (AEBP1). The colocalization percentages are calculated by the ratio of EVs with colocalized signals between AF555 and AF647 to all captured EVs (AF555) (Colocalization %, right graph). The result shows significant increases in AEBP1 -positive tEVs from the olaparib-resistant cell lines.

FIG. 9C is a series of bar graphs that show tEVs captured by anti-EpCAM and anti-CD24 immobilized on the nanosensor surface (left, orange bars). tEVs were isolated from paired cell lines (UWB1.289 PA, OVCA429 PA, PEO-1) and their resistant models for olapanb (UWB1.289 PA REOL, OVCA429 PA REOL, PEO-1 REOL). The isolated tEVs were labeled by TFP-AF555. The captured EVs were then immunolabeled by INHBA followed by AF488-conjugated secondary antibody. The AF488 signals are shown in blue bars (Marker Channel, middle). EV counts with co-localized signals of AF555 (EVs) and AF488 (INHBA). The colocalization percentages are calculated by the ratio of EVs with colocalized signals between AF555 and AF488 to all captured EVs (AF555). The result shows significant increases in INHBA-positive tEVs from the olaparib-resistant cell lines.

FIGs. 10A-10D are a series of bar graphs that show the results of testing EVs for AEBP1 in tEVs when patients were treated with olaparib (left graphs 10A and 10B) vs when different patients were treated with other drugs (right graphs IOC and 10D). P54, 60, 61, and 62 are plasma samples from ovarian cancer patients who received PARPi treatments and whose tumors progressed. P55, 57, 58, and 59 are samples from ovarian cancer patients who received other drugs than PARPi and whose tumors progressed. FIGs. 10A and IOC (top graphs) show the captured tEV counts on the substrate by anti- EpCAM or anti-CD24 (TFP-AF555). FIGs. 10B and 10D (bottom graphs) show colocalization percentages for AEBP1. The percentage was the ratio of AEBP1 -positive tEVs to all tEVs captured.

FIGs. 11A-11D are a series of bar graphs that show the results of testing EVs for INHBA in tEVs when patients were treated with olaparib (left graphs 11A and 11B) vs when different patients were treated with other drugs (right graphs 11C and 11D). P54, 60, 61, and 62 are plasma samples from ovarian cancer patients who received PARPi treatments and whose tumors progressed. P55, 57, 58, and 59 are samples from ovarian cancer patients who received other drugs than PARPi drugs and whose tumors progressed. FIGs. 11A and 11C (top graphs) show the captured tEV counts on the substrate by anti-EpCAM or anti-CD24 (TFP-AF555). FIGs. 11B and 11D (bottom graphs) show colocalization percentages for INHBA. The percentage was the ratio of INHBA- positive tEVs to all tEVs captured.

FIGs. 12A-12D are a series of bar graphs that show the results of testing tEVs for INHBA from plasma samples obtained at two time points: Tl) before treatment and T2) 2-6 weeks after initial PARPi treatment in three patients. These patients are currently undergoing olaparib treatment and showing positive clinical responses. FIG. 12A shows tEV counts captured on the nanosensor functionalized with capture antibodies of anti- EpCAM and anti-CD24. FIG. 12B shows INHBA counts after immunolabeling. FIG. 12C shows colocalized signals between captured tEVs (AF555) and INHBA signals (AF647). FIG. 12D shows the colocalized percentages defined by the ratio of colocalized tEVs to all captured tEVs.

FIGs. 13A-13D are a series of bar graphs that show the results of testing tEVs for AEBP1 from plasma samples obtained at two time points: Tl) before treatment and T2) 2-6 weeks after initial PARPi treatment in three patients. These patients are currently undergoing olaparib treatment and showing positive clinical responses. FIG. 13A shows tEV counts captured on the nanosensor functionalized with capture antibodies of anti- EpCAM and anti-CD24. FIG. 13B shows AEBP1 counts after immunolabeling. FIG. 13C shows colocalized signals between captured tEVs (AF555) and AEBP1 signals (AF499). FIG. 13D shows the colocalized percentages defined by the ratio of colocalized tEVs to all captured tEVs. DETAILED DESCRIPTION

As described herein, the use of circulating biomarkers (e.g., circulating tumor cells, soluble proteins, and cell-free DNAs) from non-invasive "liquid biopsies" offers accessible and sequential probing of molecular information from primary and metastatic tumors, marking a new era in cancer management. EVs are membrane-enclosed vesicles actively released from cells, carrying a diverse array of biomolecules like transmembrane proteins, intracellular proteins, and RNAs originating from their parent cells. Studies increasingly indicate that tumor cells secrete EVs at elevated rates compared to normal cells, and their presence has been identified in the blood of ovarian cancer patients (Jo et al., Adv Sci (Weinh) 2023; 10(27):e2301930. doi: 10.1002/advs.202301930; Yokoi et al., Science Advances; 2023; Vol 9, Issue 27, DOI: 10.1126/sciadv.ade6958; Zhang et al., BMC Cancer, 2019;19(l): 1095. doi: 10.1186/sl2885-019-6176-l). Specifically, analyses of tumor-derived EVs (tEVs) offer minimally invasive repeated sampling and provide relatively unbiased insights into the entire tumor, less affected by sample scarcity or intratumoral heterogeneity.

Despite their promising potential, the clinical application of EVs as tools for monitoring drug response is hindered by several technical challenges. Assays for detecting and quantifying EVs for specific markers often require large sample volumes and extensive processing, limiting sensitivity at the single EV level. The limited number of studies on EV-cargo associated with olaparib resistance complicates the identification of robust biomarker candidates. Moreover, the identification of cell-specific EVs and the interrogation of drug-resistance markers within specific EV subpopulations necessitate multiplexed analysis, ideally at a single EV resolution (Shao H. et al., Chem Rev., 2018; 118(4): 1917-1950, doi: 10.1021/acs.chemrev.7b00534). The methods descnbed herein have overcome these challenges in a manner that enables the successful integration of these tests into clinical practice.

General Methodology

The methods described herein include isolating tEVs from a biological sample, e.g., a whole blood sample, plasma sample, ascites, or urine, from a patient at two or more time points, e.g., before treatment, during treatment, and after treatment, for a hematologic malignancy or a solid tumor. To collect tEVs from cell cultures, one can obtain a medium that had been in contact with the cells for a sufficient time, e.g., 6, 12, 18, 24, 30, 36, 42, or 48 hours. The conditioned medium can be collected through a cell strainer, e.g., a 40 pm nylon strainer (e.g., from Thermo Fisher) and filtered, e.g., through a 0.2 m membrane filter (e.g., from Millipore Sigma). The conditioned medium can be concentrated, e.g., with a Centricon® Plus-70 Centrifugal Filter (e.g., MWCO = 10 kDa, Millipore Sigma) and centrifuged, e.g., at 3,500 g for 30 minutes, at a low temperature, e.g., at 4°C.

Following concentration, EVs are isolated from the medium, e.g., by passing the medium through a size exclusion chromatography column (SEC). For example, an SEC column can be prepared with Sepharose® CL-4B (GE Healthcare) based on known protocols (see, e.g., Van Deun et al., Adv. Biosyst. 2020, 4, 1900310). Briefly, an 11 pm pore-sized nylon membrane (NY1102500, Millipore Sigma) can be placed on the bottom of a syringe, e.g., a 10 ml syringe (BD Biosciences). The syringe can be stacked with 10 mL of Sepharose® and triple-washed with Phosphate Buffered Saline (PBS). Then, a bile sample is applied, and the 4th and 5th fractions (1 fraction = 1 mL) are collected to isolate the EVs. The collected sample can be concentrated using an Amicon Ultra-2 Centrifugal Filter (MWCO = 10 kDa, Millipore Sigma) and centrifuged at 3500 x g for 30 minutes at 4 °C. The isolated EVs can be resuspended in PBS.

Similar procedures can be applied to plasma samples, initially involving centrifugation to eliminate cell debris. EV isolation from plasma samples employs a modified SEC column, known as dual-mode chromatography (DMC)(see. e.g., Van Deun 2020). This modification in the column is useful for clinical samples, aiding in eliminating plasma lipoprotein particles (LPPs) that might overlap in size with EVs, potentially causing artifacts and leading to an overestimation of EV counts.

Following tEV isolation, the EVs are labeled, e.g., with a reporter group such as a fluorophore, such as TFP-AF555 using the published protocols (see, e.g., Ferguson et al., Sci. Adv. 2022, 8, eabm3453.). Briefly, EVs are mixed with a reporter group, e.g., 0.2 pl of TFP-AF555, followed by an incubation for a sufficient time to bind to and label the EVs, e.g., about one hour of incubation. The labeled EVs are filtered, e.g., with a 40 K MWCO column (Thermo Fisher) to remove the remaining dye and can then be diluted in PBS before use in the assays described herein. The TFP labeling is used to define the "EV signal" from the signal of the target of interest, e.g., specific surface protein biomarkers that are present on EVs, but not on other types of cells. This labeling is based on sulfo- NHS esters that bind to amine groups, which provides universal labeling of isolated EVs.

The fluorescence-amplified extracellular vesicle sensing (FLEX) technology can be employed for single EV isolation and detection (Jeong et al., Adv. Sci. 2023, 2205148). The FLEX chip, comprising periodic gold nanowell arrays, enhances immunofluorescence signals of captured EVs, thereby increasing sensitivity. After chip surface preparation, EVs are loaded and allowed to attach to the surface of the chip. Subsequently, a blocking and fixation step is performed to reduce non-specific interactions with antibodies and presen e the structural and antigenic integrity of the target of interest, e.g., the surface tumor drug-resistance biomarkers. Specific primary and secondary antibodies are then used to detect the EVs from a specific tumor type, and then to detect specific tumor resistance biomarkers. The level of tEV biomarkers that show a correlation with drug resistance or target antigen expression on tumor cells are measured and the changes in tumor drug-resistance biomarkers in all EV subpopulations and in tEv subpopulations are analyzed.

Images can be captured using a Zeiss upright automated epifluorescence microscope, e.g., with a magnification of 40x. Each fluorescence channel can be exposed for a sufficient time, e.g., 5 seconds. Image analysis can be conducted using ImageJ and custom-built Jupyter Notebook code. The analysis can include background intensity subtraction to increase the difference between real/noise signals. The ImageJ Comdet® plugin can be employed to detect EV locations from the AF555 channel.

This use of longitudinal changes in tEV tumor drug-resistance biomarkers can predict treatment efficacy and identify patients who develop resistance to a given therapy. The accuracy of the analysis is significantly improved by quantitative measurements of multiplexed marker analysis in single EVs. The methods described herein have been validated with in vitro samples for resistance to PARP inhibitors, and have been tested using serially collected human plasma samples of ovanan cancer patients.

In the examples described below we identified non-invasive EV-cargo markers to detect olaparib resistance acquisition. The findings revealed substantial insights. Firstly, we established olaparib-resistant cell lines that revealed alterations in doubling times, cell cycle distributions, and DNA repair responses, providing a robust foundation for understanding resistance mechanisms. In addition, proteomic analysis unveiled differential protein enrichment, delineating potential processes contributing to olaparib resistance. We then identified five potential markers based on their association with ovarian cancer pathophysiology and PARP mechanisms.

Preliminary pre-clinical and clinical evaluations using nanoplasmonic platforms show NHBA and AEBP1 as promising markers. Furthermore, a suggested increase in INHBA-positive tEV counts in patients treated with PARPi whose tumors progressed reinforced the potential predictor of olaparib sensitivity power of INHBA. These encouraging results indicate the feasibility of using EV-cargo markers to detect olaparib resistance.

Extracellular Vesicles (EVs)

Extracellular vesicles (EVs) are lipid-based microparticles, nanoparticle, or protein-rich aggregates present in a sample (e.g., a biological fluid) obtained from a subject. EVs also include membrane vesicles secreted from cell surfaces (ectosomes), internal stores (exosomes), cancer cells (oncosomes), or released as a result of apoptosis and cell death. In addition to lipid membranes, depending on their cell or tissue of origin, EVs can include additional components such as lipoproteins, proteins, nucleic acids, phospholipids, amphipathic lipids, gangliosides and other particles contained within the lipid membrane or encapsulated by the EVs. EVs can also be called nanovesicles.

All cells likely release, secrete, or shed EVs, making them useful clinical diagnostic and therapeutic targets for a range of diseases. Non-limiting examples of normal or cancer cell types that can release EVs include liver cells (e.g., hepatocytes), lung cells, spleen cells, pancreas cells, colon cells, skin cells, bladder cells, eye cells, brain cells, esophagus cells, cells of the head, cells of the neck, cells of the ovary, cells of the testes, prostate cells, placenta cells, epithelial cells, endothelial cells, adipocyte cells, kidney cells, heart cells, muscle cells, blood cells (e.g., white blood cells, platelets), and combinations of the foregoing. Because EVs are involved in cell-cell communication, their characterization casts light upon their role in normal physiology and pathology. EVs in biological fluids including saliva, urine, plasma, and serum can be interrogated as biomarkers of any number of cancers described herein. An EV enriched or isolated based on having particular tumor biomarkers is referred to herein as a tumor-derived EV (tEVs).

In some embodiments, an EV is between about 20 nm to about 200 nm in diameter. Individual EVs have -1/10,000 the surface area and - 1/1,000,000 the volume of a whole cell and are therefore difficult to detect using single cell analysis tools, including conventional flow cytometry. As a result, most proteomic and genomic analysis is performed in bulk on thousands or millions of EVs. However, EVs in biofluids come from many different cell types, and from different locations from within the cell (exosomes secreted from intracellular multi-vesicular bodies, ectosomes/microvesicles shed from the plasma membrane surface, membrane fragments released as a result of cell apoptosis, necrosis, etc.). Thus, in a bulk analysis, the signature from tumor EVs may be lost in the background of vesicles from other sources, and methods of enriching tEVs help capture a more robust tEV picture.

EVs represent new opportunities as circulating cancer biomarkers. These cell- derived membrane-bound vesicles contain protein and nucleic acid cargo, providing a representative “snapshot” of the content of the secreting cells. Large abundance and ubiquitous presence of tumor-derived EVs (tEVs) in bodily fluids (e.g., blood, urine) have shown the potential use of EVs as readily accessible biomarkers. Namely, tumor-derived EV (tEV) analyses can be minimally invasive for repeated sampling and afford relatively unbiased readouts of the entire tumor, less affected by the scarcity of the samples or intratumoral heterogeneity. This suggests that the methods described herein have particular utility for longitudinal disease monitoring and early detection of relapse. Previous studies showed that both the amount and molecular profiles of tEVs were shown to correlate with tumor burden. Therefore, EVs can function as a novel biomarker for liquid biopsy in personalized medicine. However, EVs are relatively new targets for analytical assays in clinics and possess unique physical and biological traits. They fall in size range much smaller than cells, but larger than proteins, and exist in a highly heterogeneous biological background. These properties impose technical difficulties, which often lead to variable findings. Furthermore, identifying cell-specific (e.g., tumor origins) EVs and interrogating drug-resistance markers within the subpopulation require multiplexed analysis, ideally in a single EV resolution.

Thus, the present disclosure provides methods of isolating and enriching tumor EV particles, for use in monitoring and/or evaluating whether tumor cells in a subject have become resistant to PARPi drugs over time.

Poly (ADP-Ribose) Polymerase (PARP) Inhibitor Resistance

PARP inhibitor resistance can arise from altered protein recruitment and trafficking patterns. Analysis of changes in the extracellular vesicle proteome upon resistance identifies proteins that have altered distribution. We have found specific proteins in the EV proteome that play critical roles in PARPi resistance to enable a better understanding of what drives resistance. Furthermore, EV analysis serves as a translatable, liquid biopsy tool to determine resistance in ovarian cancer patients.

Poly(ADP-ribose) (PARP) inhibitors (PARPi) such as olaparib, rucaparib, niraparib, and talazopanb, are now commonly used as second-line and maintenance therapy in ovarian and other cancers, with many patients showing complete remission. However, a significant portion of patients on maintenance PARPi therapy eventually develop resistance, and in large part due to the complexity of the DNA damage response, tumor resistance to PARPi is more prominent than anticipated. Some have suggested that alternative mechanisms of DNA damage repair, including reversion of the BRCA mutations to enable HR19, arise under PARPi treatment, rendering patients resistant to the otherwise effective drug. However, the majority of patients still develop PARPi resistance through unknown mechanisms, demonstrating that PARPi resistance remains largely unsolved.

Functionally, PARP is an enzyme that, upon activation through binding DNA, converts NAD+ into poly(ADP-ribose), generating a highly negative bio-polymer that functions to recruit hundreds of other proteins to sites of DNA damage and replication. PARPI is an abundant and highly productive enzyme. The over-activation of PARPI leads to cellular death via reducing the NAD(H) content of the cell, referred to as parthanatos (Fatokun et al., Br J Pharmacol, 2014; 171(8) 2000-2016. doi: 10.1111/bph. 12416). Inhibition of PARP activity with PARPi drugs prevents protein recruitment and efficient repair of single-stranded DNA damage, which generates doublestranded breaks. When cells are deficient in repairing double-stranded breaks (such as cells harboring BRCA mutations), PARP inhibition causes cellular death. Curiously, knocking out PARP is not as lethal as inhibiting PARP, which led to the observation that PARPi “traps” PARP onto DNA. However, the mechanism of PARP trapping remained mysterious. We recently found that PARP trapping is a kinetic phenomenon where the lack of protein recruitment in the absence of PARP activity' allows PARP to rebind DNA4 - PARP does not get replaced because DNA binding proteins have not been recruited to compete with PARP. However, we subsequently found that PARPi resistance can arise from PARP activity-independent recruitment of proteins. Here, the DNA damage response protein RPA1 showed DNA damage recruitment patterns (measured by image correlation spectroscopy) that strongly correlated to cellular sensitivity to PARPi across nine cell lines (Y amulla et al., Cell Rep., 2020; 32(9) 108086. doi: 10.1016/j.celrep.2020. 108086). These results suggest that altered protein trafficking and distribution impact resistance.

EVs Present a Pathway to Characterize PARPi Resistance

EVs are nano-sized, membrane-enclosed vesicles actively shed by cells. EVs carry' a set of biomolecules (e g., transmembrane and intracellular proteins, RNAs) from their originating cells, which can serve as cellular surrogates (Im et al., Nat Biotechnol. 2014; 32(5) 490-495. doi: 10.1038/nbt.2886; Ramirez-Garrastacho et al., Br J Cancer. 2022; 126(3) 331-350. doi: 10.1038/s41416-021-01610-8; Shao et al., Nat Med. 2012; 18(12) 1835-1840. doi:10.1038/nm.2994; Shao et al., Nat Commun. 2015; 6 6999. doi: 10.1038/ncomms7999). It is increasingly clear that EVs are secreted by tumor cells at higher rates than normal cells and can be identified in the blood of patients with cancer, e.g., ovarian cancer (Jo et al., Adv Sci (Weinh). 2023; e2301930. doi: 10.1002/advs.202301930; Yokoi et al., Sci Adv. 2023; 9(27) eade6958. doi: 10.1126/sciadv.ade6958; Zhang et al., Nat Biomed Eng. 2019; 3(6) 438-451. doi: 10.1038/s41551-019-0356-9). Namely, tumor-derived EV (tEV) analyses can be minimally invasive for repeated sampling and afford relatively unbiased readouts of the entire tumor, less affected by the scarcity of the samples or intratumoral heterogeneity. This suggests that tumor-derived EVs (tEVs) have particular utility for longitudinal disease monitoring and early detection of relapse (Lane et al., Clin Transl Med. 2018; 7(1) 14. doi:10. 1186/s40169-018-0192-7).

Our previous studies showed that both the amount and molecular profiles of tEVs were shown to correlate with tumor burden as well as treatment efficacy (see, e.g., Yang et al., Sci Transl Med. 2017; 9(391) eaal3226. doi: 10.1126/scitranslmed.aal3226). As discussed herein, for ovarian cancer, we found that the EV proteome (proteins found in EVs) changes upon cellular development of PARPi resistance, and these results suggest that EVs carry protein markers of PARPi resistance. In particular, the tumor drugresistance biomarkers for PARPi drugs include inhibin, beta A (INHBA), which is a protein that in humans is encoded by the INHBA gene. INHBA is a subunit of both activin and inhibin, two closely related glycoproteins with opposing biological effects. Another tumor drug-resistance biomarker is Adipocyte Enhancer-Binding Protein 1 (AEBP1), which is a protein that in humans is encoded by the AEBP1 gene. AEBP1 is a member of carboxypeptidase A protein family, which may function as a transcriptional repressor and play a role in adipogenesis and smooth muscle cell differentiation as well as in wound healing and abdominal wall development. Overexpression of this gene is associated with glioblastoma.

In other implementations, the PARPi drug resistance biomarkers include SOD3, CD44, MMP2, and TIMP1. implementations, the PARPi drug resistance biomarkers include SOD3, CD44, MMP2, and TIMP1. Further PARPi drug resistance biomarkers that have shown potential include reversion mutations in BRCA1/2, which can restore the functionality of BRCA genes, conferring resistance specifically in tumors originally sensitive due to BRCA mutations (see, e.g., nature.com/articles/nrc.2015.21). Loss of 53BP1 is associated with resistance in BRC Al -deficient tumors by facilitating alternative DNA repair pathways (see, pubmed.ncbi.nlm.nih.gov/23103855/). Upregulation of P- Gly coprotein can lead to decreased intracellular drug concentrations, affecting multiple tumor types treated with PARPi (see, nature.eom/articles/s41568-018-0005-8). Amplification of Cyclin-Dependent Kinase 12 (CDK12) has been implicated in PARPi resistance and is studied primarily in ovarian and breast cancers but may also affect other cancers (see, pubmed.ncbi.nlm.nih.gov/ 27880910/). Overexpression of RAD51 can confer resistance across various cancer types by enhancing DNA repair capabilities (see, pubmed.ncbi.nlm.nih.gov/28976962/). Loss ofPTEN, especially relevant in prostate cancer, can promote DNA repair, conferring resistance to PARPi (see, pubmed.ncbi.nlm.nih.gov/20049735/). These biomarkers vary in their tumor specificity. Reversion mutations in BRCA1/2 are highly specific to cancers that have initial BRCA- related sensitivity, whereas biomarkers like RAD51 overexpression and P-Glycoprotein upregulation are less specific and could play a role in resistance across multiple tumor types treated with PARPi

Challenges with Quantitative, Multiplexed Single EV Analysis

Despite the promises, the major bottleneck of clinical applications of EV analysis is the development of reliable, sensitive, and quantitative EV analysis. Ideally, such analysis should be useful to analyze multiple biomarkers in single EVs. This is because i) almost all types of cells shed EVs as background; ii) tEV amounts can be minuscule in small sizes of primary tumors; iii) not all tEVs contain tumor biomarkers. Thus, molecular characterization of single EVs is technically challenging. Most EVs are small vesicles (<200 nm) with limited numbers of epitopes and surface areas for labeling (i.e., weak detectable signals), which often requires sophisticated multi-step signal amplification strategies, such as DNA barcodes or enzymatic signal amplification (digital ELISA). Flow cytometry often underestimates EV counts because many small vesicles could be missed due to their weak light scattering, or a swarm of vesicles could be counted as a single event. More importantly, many of these methods are suboptimal in detecting and quantifying very rare tumor-derived EVs in excessive background EVs and particles. Various single EV methods are summarized in Table 1 below. Table 1. Single EV Detection Technologies

All of these single EV detection methods can be used in the methods disclosed herein, but the FLEX and nano-plasmonic exosome (nPLEX) methods discussed herein may be the most useful.

Methods of Monitoring Drug Resistance Using tEVs

The present methods include isolating particular EV populations (e.g., tumor- derived EVs (tEVs), e.g., ovarian cancer-derived tEVs) in a subject being treated for cancer and measuring the tEVs expression of tumor drug-resistance biomarkers over time and further administering PARPi drug therapy informed by the relative levels of tumor drug-resistance biomarker-positive (e.g., INHBA- and AEBP1 -positive) tEVs over time.

A subject is an individual (e.g., a mammal such as a human) having or suspected of having cancer, e.g., a patient diagnosed with cancer. In some embodiments, the subject can be receiving PARPi drug therapy, and/or another type of cancer treatment (e.g., radiation, surgery). A sample is typically obtained from the subject, or the sample can be from a cell culture, e.g., containing cells from a subject. A sample can include, but is not limited to, cells, lysed cells, cellular extracts, nuclear extracts, extracellular fluid, media in which cells (e.g., cancer cells from the subject) are cultured, blood, plasma, serum, gastrointestinal secretions, homogenates of tissues or tumors, ascites, synovial fluid, feces, saliva, sputum, cyst fluid, amniotic fluid, cerebrospinal fluid, peritoneal fluid, lung lavage fluid, semen, lymphatic fluid, tears, and prostatic fluid.

In some implementations, the sample is obtained from a subject at multiple time points, e.g., at least two time points. In other implementations, a single sample can be compared to a reference, e.g., a similar sample from an individual known to be healthy and/or cancer-free or to have cancer with a good response to the PARPi drug. In some embodiments, the sample from the subject can be enriched for tEVs, e.g., based on the presence of EV tumor biomarkers, but this is not required. For example, for analyzing cultured cells, it may not be required that the tEVs are isolated from other EVs in the sample. However, for clinical samples, one may wish to include the step of first isolated EVs from the sample, and then isolating tEVs from the larger EV population.

An EV tumor biomarker profile can indicate the origin of a cancer or the type of cancer cells found in a sample from a subject. For example, MUC1, HER2, EGFR, and EpCAM are four biomarkers that can be used to identify breast cancer cells in a subject. Many EVs secreted by these breast cancer cells also contain these four tumor markers. Therefore, monitoring EVs that comprise one or more of MUC1, HER2, EGFR, and EpCAM, e.g., tEVs, in a subject can give information regarding a subject s breast cancer, including development of drug resistance.

Other EV tumor biomarkers include: EpCAM, miRNA-21, and CD24 for ovarian cancer); EpCAM, EGFR, MUC1, WNT2, and GPC1 for pancreatic ductal adenocarcinoma (PDAC)); EGFR and EGFRvIII are tumor biomarkers of glioblastoma (GBM)); and EpCAM, EGFR, and MUC1 for cholangiocarcinoma); see, e.g., Im et al., Label-free detection and molecular profiling of exosomes with a nano-plasmonic sensor, Nat. Biotech. 2014; Yang, et al, Multiparametric plasma EV profiling facilitates diagnosis of pancreatic malignancy. Sci. Transl. Med. 9, eaal3226 (2017); Min et al., Plasmon- enhanced biosensing for multiplexed profiling of extracellular vesicles, A dv. Bio., 2000003 (2020); Jeong, et al., Plasmon-enhanced single extracellular vesicle analysis for cholangiocarcinoma diagnosis, Adv. Sci., 2205148 (2023). Hong, et al., CRISPR/Casl3a- Based MicroRNA Detection in Tumor-Denved Extracellular Vesicles, Adv. Sci., 2023.

Any of these tumor markers can be used as EV tumor biomarkers in the methods described herein. Other EV tumor biomarkers including miRNA and other non-coding RNAs can be found in the art and readily appreciated by the skilled artisan. See, e.g., Huang, et al., Non-coding RNA derived from extracellular vesicles in cancer immune escape: Biological functions and potential clinical applications, Cancer Lett., 2021.

In some embodiments, the methods can include using antibodies or antigen binding portions thereof that bind to selected EV tumor biomarkers corresponding to a particular type of cancer to identify or enrich tEVs for further analysis. For example, the antibodies can be capture antibodies that are attached to a substrate (e.g., a plate, well, or beads). A sample from a subject, e.g., a sample comprising a population of EVs (optionally EVs obtained from a biofluid such as blood, serum, or plasma) can then be applied to the substrate, wherein the antibodies or antigen binding portions thereof that bind to the selected EV tumor biomarkers capture and enrich the EV population for tEVs having the specified EV tumor biomarkers.

Tumor drug-resistance biomarkers can then be evaluated in the tEVs, e.g., optionally using antibodies that specifically bind the tumor drug-resistance biomarkers, to determine a level of drug-resistance biomarker for that sample and subj ect. In some embodiments, the antibodies or antigen binding portions thereof that bind to the selected EV tumor drug-resistance biomarkers can be applied to a sample, wherein the sample has been previously enriched for tEVs. One method for enriching a sample for EVs can include subjecting the sample to a plasmon-enhanced EV assay. For example, the sample can be applied to a 3D plasmonic nanostructure composed of spherical Au nanoparticles on 3D Au nanopillars (NPOP) substrate, wherein EVs are captured by the NPOP substrate. See Park, et al. Self-assembly of nanoparticle-spiked pillar arrays for plasmonic biosensing, Adv. Fund. Mater., 1904257 (2019).

In some embodiments, the antibodies or antigen binding portions thereof that specifically bind to the selected EV tumor biomarkers can be applied as free antibodies to the EV sample, wherein the antibodies or antigen binding portions thereof that bind to the selected EV tumor biomarkers can be labeled (e.g., fluorescently labeled) or wherein the antibodies or antigen binding portions thereof that bind to the selected EV tumor biomarkers can be detected with a secondary antibody. In some embodiments the antibodies or antigen binding portions thereof that bind to the selected EV tumor markers can be applied before, after, concurrently with the antibodies or antigen binding portions thereof that bind to selected tumor drug-resistance biomarker(s).

In some embodiments, EVs from a sample can be enriched using a plasmon- enhanced EV capture method. In some embodiments, the plasmon-enhanced EV capture method includes EV capture using any substrate, e, g., plain substrate, nanostructures, beads, or other materials. In some embodiments, the plasmon-enhanced EV capture method includes EV capture using an NPOP substrate, wherein in some embodiments the NPOP substrate can be constructed and/or functionalized according to the methods described in the examples. EVs that have been enriched by isolation on an NPOP substrate can be probed for expression of EV tumor biomarker(s) and/or tumor drugresistance biomarker(s). Antibodies to EV tumor biomarkers can be applied to the EV- enriched sample, wherein the antibodies to EV tumor biomarker(s) can be labeled (e.g., fluorescently labeled) or wherein secondary antibodies can be used to detect the antibodies to EV tumor biomarker(s).

Antibodies to tumor drug-resistance biomarker(s) can be applied to the EV- enriched sample, wherein the antibodies to the tumor drug-resistance biomarker(s) can be labeled (e.g., fluorescently labeled) or wherein secondary antibodies can be used to detect the antibodies to the tumor drug-resistance biomarker(s). In some embodiments, the antibodies to EV tumor biomarker(s) and the antibodies to the tumor drug-resistance biomarker(s) can be applied to the sample and/or the EV-enriched sample at the same time. In some embodiments, the antibodies to the EV tumor biomarker(s) are applied to the sample and/or EV-enriched sample prior to when the antibodies to the tumor drugresistance biomarker(s) are applied to the sample and/or EV-enriched sample. In some embodiments, the antibodies to the EV tumor biomarker(s) are applied to the sample and/or EV-enriched sample after the antibodies to the tumor drug-resistance biomarker(s) are applied to the sample and/or EV-enriched sample.

In some embodiments, enriching EVs (e.g., tEVs) and probing the tEVs for levels of tumor drug-resistance biomarker(s) can be carried out at multiple time points (e.g., over time or longitudinally). For example, enriching EVs (e.g., tEVs) and probing the tEVs for levels of drug-resistance biomarkers can be carried out at one, two, three, four, five, or more time points. In some embodiments, the level (as determined by antibody detection) of tumor drug-resistance biomarker(s) at a first time point can be used to determine the relative level of the tumor drug-resistance biomarker(s) at a second time point by comparing the level of tumor drug-resistance biomarker(s) signal at the second time point to the tumor drug-resistance biomarker(s) signal at the first time point and noting an increase or decrease of the level of the tumor drug-resistance biomarker(s) signal. Similarly, the level of the tumor drug-resistance biomarker(s) signal at a third (or fourth, or fifth, etc.) time point can be compared to the level of drug-resistance biomarker(s) signal at the first time point, or the level of the drug-resistance biomarker(s) signal at the third (or fourth, or fifth, etc.) time point can be compared to the level of drug-resistance biomarker(s) signal at any previous time point to analyze whether there are any trends in drug-resistance biomarker(s) signal over time.

In some embodiments, a relative increase in levels of the tumor drug-resistance biomarker(s)-positive (e.g., INHBA and AEBP1) tEVs over a previous level of tumor drug-resistance biomarker(s)-positive (e.g., INHBA- and/or AEBP1 -positive) tEVs indicates cancer cells in the subject are in the process of becoming, or have become, resistant to PARPi drugs, for example the PARPi drug used to treat the subject’s cancer. In some embodiments, a relative decrease or no significant change in levels of drugresistance biomarker(s)-positive (e.g., INHBA- and/or AEBP1 -positive) tEVs over the previous level of the tumor drug-resistance biomarker(s)-positive tEVs indicates cancer cells in the subject have not become resistant to the PARPi drug (e.g., used to treat the subject’s cancer).

The relative levels of tumor drug-resistance biomarker(s) and biomarker(s)- positive tEVs thus can be used to determine whether the subject receives additional PARPi drug treatments comprising the same drug used to treat the subject’s cancer between the previously evaluated time points or receives a different treatment using a different chemotherapeutic agent (or other treatment modality, such as immunotherapy, radiotherapy, or surgical resection).

In some embodiments, a relative increase in tumor drug-resistance biomarker(s) or biomarker(s)-positive (e.g., INHBA and/or AEBP1) tEVs over a previous level of drugresistance biomarker(s)-positive tEVs indicates whether the subject should continue to receive treatment with the same PARPi drug. In some embodiments, a relative increase in drug-resistance biomarker(s)-positive tEVs over the previous level of drug-resistance biomarker(s)-positive tEVs indicates the subject should not receive further treatment with the same PARPi drug. In some embodiments, a relative increase in drug-resistance biomarker(s)-positive EVs over the previous level of drug-resistance biomarker(s)- positive tEVs indicates the subject should not receive treatment with any further PARPi drug.

In some embodiments, the subject only receives further PARPi drug treatments with the same drug if there is a decrease or no change in the tumor drug-resistance biomarker(s)-positive tEVs over the previous (or any previous) level of drug-resistance biomarker(s)-positive tEVs. In some embodiments, the subject receives further PARPi drug treatments with the same drug only if there is not an increase in drug-resistance biomarker(s)-positive tEVs over the previous (or any previous) level of drug-resistance biomarker(s)-positive tEVs. In these conditions, administration of further treatments with the same drug is dependent on the relative level of drug-resistance biomarker(s)-positive tEVs over the previous (or any previous) level of drug-resistance biomarker(s)-positive tEVs.

As used herein, the terms “cancer,” “tumor” or “tumor tissue” has the meaning as understood by one skilled in the art. A cancer, tumor, or tumor tissue can include tumor cells that are neoplastic cells with abnormal growth properties. Tumors, tumor tissue, and tumor cells can be benign or malignant. Cancer can include primary malignant cells or tumors (e.g., those whose cells have not migrated to sites in the subject’s body other than the site of the original malignancy or tumor) and secondary malignant cells or tumors (e.g., those arising from metastasis, the migration of malignant cells or tumor cells to secondary sites that are different from the site of the original tumor).

Examples of cancer include, but are not limited to, carcinoma, lymphoma, blastoma, sarcoma, and leukemia or lymphoid malignancies. Additional examples of such cancers are noted below and include: squamous cell cancer (e.g., epithelial squamous cell cancer), lung cancer including small-cell lung cancer, non-small cell lung cancer, adenocarcinoma of the lung and squamous carcinoma of the lung, cancer of the peritoneum, hepatocellular cancer, gastric or stomach cancer including gastrointestinal cancer, pancreatic cancer, glioblastoma, cervical cancer, ovarian cancer, liver cancer, cholangiocarcinoma, bladder cancer, hepatoma, breast cancer, colon cancer, rectal cancer, colorectal cancer, endometrial cancer or uterine carcinoma, salivary gland carcinoma, kidney or renal cancer, prostate cancer, vulvar cancer, thyroid cancer, hepatic carcinoma, anal carcinoma, penile carcinoma, as well as head and neck cancer.

One of the benefits of the longitudinal monitoring of drug-resistance of the currently described methods is that drug-resistance in a subject can be detected prior to an observable increase in size of the subject’s cancer (e.g., tumor). With an early detection of drug-resistance, the longitudinal monitoring of drug-resistance of the currently described methods can terminate the toxic treatment early to reduce side effects or minimize unnecessary treatment. The efficacy of the longitudinal monitoring of drug-resistance of the currently described methods for predicting drug-resistance can be at least 95%. Other benefits of the longitudinal monitoring of drug-resistance of the currently described methods include predicting drug treatment efficacy, minimizing the detection of residual diseases, and facilitating early detection of disease recurrence.

Nanoplasmonic Technology for Single EV Analyses

Plasmonic EV sensing platforms, named nPLEX (nano-plasmonic exosome), that can rapidly detect and molecularly profile tumor-derived EVs in clinical samples have been described (see, e.g., US Patent No. 10,557,847B2). We further advanced the technology for multiplexed single EV analysis (see, e.g., US Patent Application Publication No. US US-2023-0160809). Termed “FLEX (fluorescence-amplified extracellular vesicle sensing),” the technology harnesses plasmonic metallic nanostructures to amplify EVs’ fluorescence signals and significantly improve the detection sensitivity down to single EVs. The signal amplification occurs in multiple colors, enabling multiplexed, multichannel imaging and detection of single EVs. The results indicate that conventional fluorescence imaging using a plain substrate detected only 10-15% of total EVs due to weak fluorescence signals of small EVs; these weak signals get amplified by using nanoplasmonic chips. The assay is simple and compatible with conventional immunostaining and imaging but does not require any additional chemical reactions to achieve enough sensitivity for single EV detection. In particular, the plasmon enhancements allow us to use near-infrared fluorophores (e.g., Cy7) that are barely used for EV imaging due to weak signals. This enabled us to develop multichannel single-EV imaging in a broader spectrum range.

The sensor chip comprises periodic gold nanowell arrays made on a Si wafer in a wafer scale, which addressed the main bottleneck of the technology with high-throughput chip production. Using the single EV analysis, we showed 1) a wide heterogeneity of EVs and their marker levels and 2) high sensitivity for rare target EVs not detected by conventional fluorescence detection due to weak signals. We also developed a new surface chemistry that significantly reduces non-specific EV binding to plasmonic gold surfaces (see. e.g., Kim et al., ACS Appl Mater Interfaces, 2022 Jun 2: 10.1021/acsami.2c07317. doi: 10.1021/acsami.2c07317; and PCT Application Publication No. WO 2023/220377), improving specific EV capture on the substrate surface (high specificity). Using the new platform, we demonstrated sensitive detection of tEVs and accurate quantification of temporal changes in tEVs, which could accurately identify patients who are not responding to therapy.

EXAMPLES

As described in the following examples, olapanb resistance models were created, and through whole proteomic and enrichment analysis, specific EV-cargo biomarkers associated with olaparib response, e.g., resistance, were identified. The expression of these biomarkers at a single EV level was investigated using nPLEX. Subsequently, these findings were validated using serial clinical samples from patients undergoing olaparib treatment with varying clinical outcomes. The inventions described herein are further described in the following examples, which do not limit the scope of the invention descnbed in the claims. The materials and methods described below have been used to generate the examples described herein.

Materials and Methods

Cell lines: Human ovarian carcinoma cell lines UWB1.289 and OVCA429, obtained from the American Type Culture Collection (ATCC), were maintained in RPMI- 1640 and Dulbecco s modified Eagle s medium, respectively. Complete growth media were supplemented with 10% FBS and 100 units/ml penicillin-streptomycin, incubated at 37°C with 5% CO2. Mycoplasma-free conditions were confirmed using the Universal Mycoplasma Detection Kit (ATCC).

Establishment of olaparib-resistant subtypes from parental (control) cell lines: Establishment of Olaparib-Resistant Subtypes: UWB1.289 REOL and OVCA429 REOL subtypes were developed by gradually exposing cells to increasing Olaparib (AZD2281, SelleckChem, USA) concentrations over 16-18 weeks. Olaparib doses were incremented stepwise up to tenfold, authenticated by STR profiling at Dana-Farber Cancer Institute, and used within six months.

Cellular Characterization: Drug Sensitivity and Cell Cycle Analysis: Drug sensitivity was evaluated using the MTT colorimetric assay. Cells were seeded in a 96- well plate and treated with olaparib (1.00E-03 -1000 pM) for 72 hours. MTT dye was added, and absorbance at 540 nm was measured using a plate reader after formazan crystal dissolution. Results represent the mean SD of three independent experiments performed in triplicates.

Cell doubling times for parental and resistant subtypes were calculated over 24, 48, 72, and 96 hours using the formula Doubling time (h) = Incubation time (h) x [1 / (Log2 (Cellsf I Celli))]. Viable cell counts were determined using Trypan Blue exclusion.

Cell lines were synchronized, treated with 10% FBS, fixed in ethanol, and stained using the Cell Cycle Analysis kit. Flow cytometry analysis (FlowJo™ Software) determined the percentage of cells in Gl, S, and G2/M phases for each parental and resistant subtype pair.

Automated Fluorescence Microscopy for DNA Damage Response Protein Analysis: To study DNA damage response, specific proteins like RAD51, yH2AX. and PCNA were assessed in parental and resistant cell subtypes. After seeding cells into a 384-well plate and fixation, primary antibodies (Table 2) were applied overnight.

Subsequently, secondary antibodies (Table 3) were used, followed by DAPI staining for 15 minutes. Images were captured using an inverted Nikon Eclipse Ti2 microscope equipped with LED lighting and a CMOS camera, allowing visualization of nuclear segmentation. The software NIS Elements version 5.30.07 managed data acquisition, enabling protein expression analysis correlating with cell phases.

Table 2 - List of Primary Antibodies

Table 3 - List of Secondary Antibodies

Extracellular Vesicle (EV) Isolation: The cell culture medium was replaced with 1% exosome-depleted FBS for 48 hours for cell line-derived EV isolation. The conditioned medium underwent filtration to remove cells and apoptotic bodies and then concentrated using centrifugal filter units (Centricon-70, 10 kDa cutoff). Centrifugation concentrated the medium, which was further processed using size exclusion chromatography (SEC) as previously reported (Jeong M. et al., Adv Sci (Weinh) . 2023 Mar;10(8):e2205148. doi: 10.1002/advs.202205148). Plasma samples were centrifuged at 2,000 * g for 3 minutes to remove cell debris, and supernatants were collected. Five hundred microliters of supernatants diluted in 500 pL of PBS were used for isolation experiments by the isolation size exclusion chromatography (EDMC) as previously reported (Woo H. et al., Thera osiics, 2022 1 an 31 ;12(5)' 1988-1998; dor 10.7150/th o 69094). NanoSight LM10 (Malvern) equipped with a 405 nm laser was used. Samples were diluted in fPBS to obtain the recommended particle concentration (25-100 particles/frame). For each test sample, three 30-sec videos were recorded (camera level, 14). Recorded videos were analyzed by NTA software (version 3.2) at a detection threshold of 3.

Proteomic Analysis: Extracellular vesicles (EVs) were isolated from UWB 1.289, OVCA429, UWB1.289 REOL, and OVCA429 REOL cells. Each EV sample was lysed in lysis buffer, precipitated using acetone, and subjected to trypsin digestion. Formic acid was added to the EV-derived peptides before centrifugation and drying. The resulting peptides were loaded onto a Cl 8 Stage Tip, eluted, and prepared for LC-MS/MS analysis. EV-derived peptides were analyzed on an LTQ-Orbitrap XL instrument coupled to an Ultimate 3000 Dionex nanoflow LC system. The RP-LC system utilized a Cap-Trap cartridge and a BioBasic C18 PicoFrit analytical column. Peptides were eluted over 180 minutes using a linear gradient of 6 to 100% mobile phase B. The LTQ-Orbitrap operated in a data-dependent mode, acquiring full MS scans followed by MS/MS scans via CID and HCD. Settings included 60,000 resolving power for MSI scans and 7,500 resolving power for MS2 scans. Mascot was used for protein identification, and Proteome Discoverer 1.2 software generated Mascot format files. Abundance differences were determined by comparing resistant subtypes with their corresponding parental EV samples.

Single EV Analysis with Nano-Plasmonic Exosome (nPLEX) Sensor: Azido- dPEG-TFP ester was reacted with DBCO Alexa Fluor fluorophore 555. Dye-PEG12-TFP conjugates labeled EV surface proteins. Typically, 2.5E10 EV was combined with 0.8 pl of Dye-PEG12-TFP. Excess Dye-PEG12-TFP was removed using two cicles of Zeba Micro Spin Desalting Columns.

The nP LEX-FL chip was incubated overnight with 10 mM biotinylated thiol-PEG in deionized water. Following this, the PEGylated chip was washed with deionized water. Neutravidin in PBS with 0.2% bovine serum albumin, at a concentration of 50 pg/mL, was applied onto the chip surface and allowed to incubate for 1 hour. The chip was then thoroughly washed to remove any unbound neutravidin. Biotinylated EVs were subsequently applied to the chip surface and incubated for 30 minutes. Afterward, the chip was washed with PBS. After the fixation/permeabilization and blocking steps primary antibodies were diluted in 1% BSA in PBST and were incubated on the chip surface for 20 minutes (Table 2). The chip was washed multiple times with PBST to remove any unbound antibodies. Fluor ophore-conjugated secondary antibodies were then diluted in 1% BSA in PBST and incubated on the chip surface for 10 minutes (Table 3). The chip underwent multiple washes with PBST to wash out unbound antibodies. Following this, the chip was mounted onto a glass slide, and a mounting solution was applied on the chip surface, covered with a covershp. The labeled EVs were imaged using a multi-channel fluorescence microscope.

Human Samples: Subjects were recruited according to an Institutional Review Board approved protocol with informed consent. A total of 6 individuals were enrolled. Blood samples were collected from patients as per routine in Massachusetts General Hospital Abdominal Imaging and Intervention suites. For longitudinal treatment response evaluation, serial ascites samples were collected from each patient (n = 6) during three distinct treatment visits. Responder and non-responder status was independently assigned by a gynecologic oncologist based on commonly used response criteria in ovarian cancer studies: 1) CA-125 based on Gynecologic Cancer Intergroup (GCIG) criteria, 2) scans based on Response Evaluation Criteria In Solid Tumors (RECIST), and /or 3) in cases where such data were not available within a week of collection, the electronic medical record for documented clinical impressions (e.g., palliative care without active therapy due to clinical decline or quality of life changes based on increased/decreased ascites accumulation).

Example 1 - Olaparib Resistance in Ovarian Cancer Cell Lines

Generation of Olaparib Resistance (REOL) in Ovarian Cancer Cell Lines We selected two ovarian cancer cell lines to establish olaparib-resistant cell lines: UWB1.289, characterized by a germline BRCA1 mutation and wild-type allele deletion, and OVCA429, carrying a P53 mutation. Treatment began with a low olaparib dose, progressively escalating over time while regularly passaging the parental cell lines (FIG. 1A). Two sublines — UWB1.289 REOL and OVCA429 REOL — were generated, showing no discernible morphological differences from their parental counterparts. Assessing inhibitory concentration (IC50) values revealed a considerable increase in olaparib dosage required to inhibit 50% of cell growth in UWB 1.289 REOL and OVCA429 REOL compared to their parental lines. Specifically, IC50 values for UWB1.289 parental and UWB1.289 REOL were 0.85 0.32 and 94.39 24.3, respectively (FIG. IB). Similarly, IC50 values for OVCA429 parental and OVCA429 REOL were 10.33 6.22 and 151.86 22.45, respectively (FIG. 1C).

Doubling Time and Cell Cycle Distribution

We extended our analysis of UWB1.289 and OVCA429 parental cell lines and their corresponding REOL subtypes, given the observed shift in IC50 after olaparib treatment. Doubling times for the REOL subtypes (UWB1.289 REOL and OVCA429 REOL) were notably longer than their parental subtypes. This difference was particularly significant for the OVCA429 cell line (FIG. 2A). To delve deeper, cell cycle analysis was performed, revealing that after 24 hours of olaparib exposure, the resistant subtypes showed a notably higher proportion of cells in the G2 phase and fewer cells in the G1 phase compared to their parental lines (FIG. 2B). Notably, this trend was amplified with higher olaparib concentrations (10 pM). suggesting an increased duration spent by the resistant cells in repairing DNA damage.

DNA Damage Response Evaluation

To evaluate DNA repair responses, we examined y-H2AX, RAD51, and PCNA expression in parental and REOL cell subtypes after exposing them to varying olaparib concentrations for 24 hours. Our findings showed heightened y-H2AX expression in all cell lines, suggesting DNA damage recognition due to olaparib treatment escalation (FIGs. 3A and 3B). This aligns with prior studies highlighting y-H2AX as a sensitive biomarker for DNA double-strand breaks (DSBs) (Rogakou et al., J. Biol. Chem, 273, 5868 (1998)). The increased y-H2AX levels in S and G2 phases potentially impact DNA replication and G2 -phase transitions by initiating repair mechanisms. RAD51, a key DNA repair factor, exhibited decreased levels with increasing olaparib doses in parental cell lines (UWB1.289 and OVCA429) but remained consistent in REOL cell lines (UWB1.289 REOL and OVCA429 REOL) across concentrations and cell cycle phases (FIGs. 3C and 3D) This sustained RAD51 presence in resistant cells suggests potential hindrance in RAD51 degradation, bolstering DNA repair mechanisms, possibly contributing to resistance through enhanced HR-mediated repair or replication fork stabilization (Su et al., Nat. Struct. Mol. Biol., 15: 1049-1058 (2008); Kolinjivadi et al., Mol. Cell, 68 (2):414-430 (2017)).

Furthermore, we observed a dose-dependent rise in PCNA expression across all cell cycle stages in REOL compared to their parental cell counterparts (FIGs. 3E and 3F). Heightened PCNA levels might augment the PCNA-RAD51 interplay, possibly facilitating RAD51's loading and function at stalled replication forks, essential for their proper restart and maintenance upon damage (Choe K.N. et al., Mol Cell., 2017 Feb 2;65(3): 380-392. doi: 10.1016/j.molcel.2016.12.020).

Characterization of EVs from Parental and REOL Cell Subtypes

Once the acquisition of resistance was confirmed in both REOL subtypes, EV fractons were isolated and further characterized for the analysis of size distribution and concentration by NTA (FIGs. 4A-4D). The population of particles from both parental and resistant cell subtypes was quite homogeneous in terms of size (range 104-135 nm). The result shows that developing resistance to olaparib does not significantly change the cell’s EV secretion and physical characteristics (size and concentration). It also indicates the EV samples are adequately prepared and isolated.

Example 2 - Whole Proteome Analysis and Resistance Biomarker Candidate Selection

Detection of Proteins Exhibiting Enrichment between Parental and Resistant Cell Subtypes

To unravel the mechanisms underlying olaparib resistance, we conducted whole- proteomic sequencing on EVs extracted from UWB 1.289 and OVCA429 cellular models, comprising both parental and REOL subtypes. Venn diagram analysis revealed the presence of 641, 800, 904, and 864 proteins in the EVs derived from UWB1.289, UWB1.289 REOL, OVCA420, and OVCA429 REOL samples, respectively (FIG. 5A). Furthermore, we identified 200 differentially enriched proteins between the UWB1.289 parental and REOL-derived EV subtypes, along with 111 differentially expressed proteins (DEPs) in the OVCA429 parental and REOL-derived EV subtypes. Subsequently, we conducted a Gene Ontology (GO) analysis to explore the interactions and functional relationships among these identified DEPs. The differential protein enrichment in EVs from the UWB1.289 parental and REOL cell models indicates a significant focus on essential processes like cell cycle regulation, proliferation, and cellular motility. This enrichment is reflected in key biological processes such as Positive Regulation Of Cell Population Proliferation (G0:0008284), Cell-Matrix Adhesion (GO: 0007160), and Regulation Of Cell Population Proliferation (G0:0042127). At the cellular level, the emphasis is observed in the Collagen-Containing Extracellular Matrix (GO: 0062023), Focal Adhesion (G0:0005925), and Vesicle (G0:0031982). Moreover, this enrichment highlights specific molecular functions like Growth Factor Activity (GO: 0008083), Receptor Ligand Activity (G0:0048018), and Calcium Ion Binding (GO: 0005509). These distinct enrichments suggest tailored coordination of cell proliferation and motility, possibly contributing to the development of olaparib resistance in the UWB1.289 cell model (FIG. 5B).

Additionally, in EVs derived from OVCA429 parental and REOL cell models, the enrichment indicates potential shifts in molecular signaling pathways and ion-binding capacities. There is a notable enrichment in biological processes like Negative Regulation Of Receptor-Mediated Endocytosis (G0:0048261), Positive Regulation Of Epithelial Cell Migration (G0:0010634), and Endodermal Cell Differentiation (G0:0035987). At the cellular level, significant changes are observed in components such as Collagen- Containing Extracellular Matrix (G0:0062023), Secretory Granule Lumen (GO: 0034774), and Vesicle (G0:0031982). Furthermore, key molecular functions like Cadherin Binding (G0:0045296), Protein Kinase C Binding (G0:0005080), and Serine- Type Peptidase Activity (G0:0008236) demonstrate alterations in molecular interactions and enzymatic activities (FIG. 5C).

We next performed a differential expression analysis, selecting the proteins commonly enriched among both cellular models (UWB 1.289 and OVCA429) as shown in the heatmap of FIG. 6, which shows 20 commonly enriched DEPs. To select potential markers to continue with further evaluations, we selected these based on previous association with ovarian cancer pathophysiology and PARP mechanisms besides expression consistency across samples. We selected an initial set of 5 biomarkers involved in biological processes, including cell signaling, extracellular modulation, and responses to oxidative stress. The comparative analysis shows certain proteins associated with drug resistance are observed in the in vitro models, which represent different resistance pathways. Further investigation beyond the 5 initially selected biomarkers may strengthen the assay’s accuracy.

Example 3 - Olparaib Resistance Marker Validation

Marker Validation on Cell Line-Derived EVs

Initially, we utilized EVs derived from cell lines to validate the whole-proteomic outcomes and refine the method conditions for optimal sensitivity and minimal background noise. As shown in the schematic of FIG. 7A, we focused on two proteins, INHBA and AEBP1, displaying the most substantial average upregulation between REOL and parental cells in both UWB 1.289 and OVCA429 lines, and we then conducted EV analysis using a nanoplasmonic platform across all four cell lines (FIGs. 7B-7C). The measurements of INHBA and AEBP1 revealed significantly elevated levels in EVs from REOL UWB1.289 and OVCA429 in comparison to parental cells. These findings strongly indicate the potential success of our experimental approach in identifying and validating EV protein biomarkers associated with olaparib response and resistance. This also shows tumor cells shed more tEVs carrying INHBA or AEBP1 into the extracellular environment. This leads to a higher amount of circulating tEVs earn ing those PARPi- resistant biomarkers when tumor cells develop resistance to PARPi drugs.

Detection of Resistant EV Markers for PARPi Treatment

We performed a preliminary test measuring INHBA levels in tEVs in plasma samples at three-time points: Tl) before treatment, T2) 4-6 weeks after initial PAPRi treatment, and T3) 3-6 months after treatments. In this test, we analyzed samples from three patients who received PARPi treatment and whose tumors progressed at T3 and another three patients who received other treatments than PARPi treatments, and their tumors progressed as well (FIGs. 8A-8C). In this analysis, we first isolated EVs using size-exclusion chromatography and labeled them using TFP-AF555. The isolated EVs were captured on the sensor surface and then immunolabeled by anti-INHBA, followed by secondary antibody -labeled by AF647. We then measured increases of INHBA- positive tEV counts at T2 and T3 compared the baseline at Tl (before treatment) for plasma samples from patients who received PARPi treatment with progression. Meanwhile the INHBA-positive EV counts did not significantly change for plasma samples from patients who received non-PARPi treatments. The result indicates the increase of INHBA is specific to resistance to PARPi treatment as a PARPi drugresistance marker, as the level did not increase in patients who received treatments other than PARPi and progressed.

Example 4 - Evaluation of Expression of Biomarkers on EVs Positive for Tumoral Biomarkers (CD24+EPCAM) in Cell Line-Derived EVs

As shown in the schematic of FIG. 9A, we evaluated the expression of the two biomarkers AEBP1 and INHBA on a preselected population of EVs positive for tumoral markers (CD24+EPCAM) in cell line-derived EVs. We also included the cell line PEO-1 and a resistant cell subtype to validate that AEBP1 and INHBA are up-regulated in cell lines that were not a part of the proteomic analysis to identify the candidate markers. FIG. 9A is based on immunocapture of tEVs using anti-EpCAM and CD24 (ovarian tEV markers) and then immunolabeled by those two markers. The results were similar to those shown in FIGs. 7A-7C, and so for in vitro samples, both methods can be used. However, for clinical samples, one might prefer to use the second method to first capture tEVs, as there are many non-tEVs in blood.

We validated whole-proteomic outcomes using EVs derived from cell lines and optimized conditions for sensitivity and minimal background noise. There was a significant difference in the expression of AEBP1 between all parental and resistant cell subtype models. However, we could see a significant increase in INHBA expression only in the UWB1.289 and OVCA429 resistant subtypes compared with their parental counterparts, and the PEO1 model did not show a different expression before and after treatment. As shown in FIGs. 9B-9C, there are no changes or correlations of CD24- or EpCAM-positive tEV counts between olaparib-resistant and parental cell lines. However, looking at the colocalization percentages, the INHBA- or AEBP1 -positive tEV counts significantly increased in olaparib-resistance cell lines. This validates the potential of the identified drug-resistance markers in assessing and predicting olaparib resistance in tumor cells.

Example 5 - Evaluation of Expression of Drug-Resistance Biomarkers on tEVs isolated from plasma of ovarian cancer patients

We analyzed human plasma samples of ovarian cancer patients who received PARPi treatment. We measured AEBP1 and INHBA levels in tEVs in plasma samples at three time points: Tl) before treatment, T2) 4-6 weeks after initial PARPi treatment, and T3) 3-6 months after treatments. In these tests, we analyzed samples from eight patients whose tumors progressed at T3 after receiving PARPi treatments (P54, P60, P61, and P62; FIGs. 10A-10B) and other treatments than PARPi (P55, P57, P58, and P59; FIGs. 10C-10D) In this analysis, we first captured tEVs using anti-EpCAM and anti-CD24 as capture antibodies immobilized on the nanosensor surface. FIGs. 10A and IOC show the captured tEV counts at Tl, T2, and T3.

After specifically capturing tEVs positive for EpCAM and/or CD24, the captured EVs were immunolabeled by AEBP1, as described in Example 4. FIGs. 10B and 10D show the percentage of colocalized tEVs (i.e., AEBP1 -positive tEVs) defined by the ratio of AEBP1 -positive tEVs to all captured tEVs. In comparing AEBP1 -positive tEV counts, we also measured increases in AEBP1 -positive EV counts at T2 and/or T3 compared to the baseline at Tl (before treatment) for plasma samples from patients who received PARPi treatment with progression. Meanwhile, the AEBP1 -positive tEV counts do not significantly change for plasma samples from patients who received non-PARPi treatments. The results validate that the use of AEBP1 -positive tEVs in circulation could indicate the resistance with PARPi therapy.

FIGs. 11A and 11C show the captured tEV counts at Tl, T2, and T3. After specifically capturing tEVs positive for EpCAM and/or CD24, the captured EVs were immunolabeled by INHBA, as described in Example 4. FIGs. 11B and 11D show the percentage of colocalized tEVs (i.e., INHBA-positive tEVs) defined by the ratio of INHBA-positive tEVs to all captured tEVs. In comparing INHBA-positive tEV counts, we measured increases in INHBA-positive tEV counts at T2 and/or T3 compared to the baseline at Tl (before treatment) for plasma samples from patients who received PARPi treatment with progression. Meanwhile, the INHBA-positive tEV counts do not significantly change for plasma samples from patients who received non-PARPi treatments. The results validate that the use of INHBA-positive tEVs in circulation could indicate resistance to PARPi therapy.

While some patients showed increased in tEV counts at T2 and/or T3 compared to Tl, the drug-resistance marker analysis is more accurate than analyzing the tEV counts, supporting our methods to assess and evaluate resistance to PARPi therapy.

Example 6 - Evaluation of Expression of Drug-Resistance Biomarkers on tEVs Isolated from Plasma of Ovarian Cancer Patients who Receive PARPi Drug Therapy and Showed Good Responses In addition to the long-term treatment patients who progressed under olaparib or other chemotherapy treatments, we measured INHBA and AEBP1 levels in tEVs from plasma isolated at two time points: Tl) before treatment and T2) 2-6 weeks after initial PARPi treatment in three patients. These patients are currently undergoing olaparib treatment and showing positive clinical responses. These additional patients were processed following the same protocol previously described. We first captured tEVs using anti-EpCAM and anti-CD24 as capture antibodies immobilized on the nanosensor surface. After specifically capturing tEVs positive for EpCAM and/or CD24, the captured EVs were immunolabeled with INHBA or AEBP1.

FIGs. 12A-12D are a series of bar graphs that show the results of testing tEVs for INHBA from plasma samples obtained at two time points: Tl) before treatment and T2) 2-6 weeks after initial PARPi treatment in three patients. These patients are currently undergoing olaparib treatment and showing positive clinical responses. FIG. 12A shows tEV counts captured on the nanosensor functionalized with capture antibodies of anti- EpCAM and anti-CD24. FIG. 12B shows INHBA counts after immunolabeling. FIG. 12C shows colocalized signals between captured tEVs (AF555) and INHBA signals (AF647). FIG. 12D shows the colocalized percentages defined by the ratio of colocalized tEVs to all captured tEVs.

FIGs. 13A-13D are a series of bar graphs that show the results of testing tEVs for AEBP1 from plasma samples obtained at two time points: Tl) before treatment and T2) 2-6 weeks after initial PARPi treatment in three patients. These patients are currently undergoing olaparib treatment and showing positive clinical responses. FIG. 13A shows tEV counts captured on the nanosensor functionalized with capture antibodies of anti- EpCAM and anti-CD24. FIG. 13B shows AEBP1 counts after immunolabeling. FIG. 13C shows colocalized signals between captured tEVs (AF555) and AEBP1 signals (AF488). FIG. 13D shows the colocalized percentages defined by the ratio of colocalized tEVs to all captured tEVs.

In the analysis, there was no change in INHBA levels for any of the patients or time points (FIGs. 12A-12D). However, we observed an increase in AEBP1 levels in one of these three patients at T2 compared to Tl (FIGs. 13A-13D). This may indicate an early acquisition of resistance that needs to be closely monitored clinically. Our observations confirm that olaparib treatment itself does not significantly increase the expression of these markers in responsive patients, contrasting with the changes observed in olaparib-resistant patients. OTHER EMBODIMENTS

Whilst the invention has been disclosed in particular embodiments, it will be understood by those skilled in the art that certain substitutions, alterations and/or omissions may be made to the embodiments without departing from the spirit of the invention. Accordingly, the foregoing description is meant to be exemplary only, and should not limit the scope of the invention. All references, scientific articles, patent publications, and any other documents cited herein are hereby incorporated by reference for the substance of their disclosure.

Claims

We claim:

1. A method of monitoring a tumor’s resistance to a poly (ADP -ribose) polymerase inhibitor (PARPi) drug administered to a subject to treat the tumor, the method comprising

(a) obtaining a sample from the subject at a first time point;

(b) isolating and quantifying extracellular vesicles (EVs) from the sample based on detection of EV biomarkers on the EVs in the sample; and

(c) determining a presence or level of a tumor drug-resistance biomarker of the EVs isolated at the first time point, wherein a presence or level of the tumor drugresistance biomarker indicates a resistance by the tumor to the PARPi drug being administered to the patient.

2. A method of monitoring a tumor’s resistance to a poly (ADP -ribose) polymerase inhibitor (PARPi) drug, the method comprising

(a) obtaining a sample containing cells from the tumor at a first time point;

(b) isolating and quantifying extracellular vesicles (EVs) from the sample based on detection of EV biomarkers on the EVs in the sample; and

(c) determining a presence or level of a tumor drug-resistance biomarker of the EVs isolated at the first time point, wherein a presence or level of the tumor drugresistance biomarker indicates a resistance by the tumor to the PARPi drug.

3. The method of claim 1 or claim 2, wherein step (b) comprises isolating tumor- derived EVs (tEVs) from the sample based on detection of tumor biomarkers on the tEVs from the sample.

4. The method of claim 3, wherein step (b) comprises first isolating EVs from the sample, and then isolating tEVs from the EVs isolate from the sample.

5. The method of any one of claims 1 to 4, wherein step (c) comprises comparing the presence or level of the tumor drug-resistance biomarker to a reference sample or reference level from a subject known to be healthy and/or cancer-free or to have cancer with a good response to the PARPi drug, wherein a presence of the tumor drug-resistance biomarker in the sample when there is no presence or low levels of the tumor drug- resistance biomarker in the reference, or when a level of the tumor drug-resistance biomarker that is higher in the sample than the reference level, indicates a resistance by the tumor to the PARPi drug.

6. The method of any one of claims 1 to 5, further comprising

(d) obtaining a sample, e.g., from the subject, at a second time point, which is after the first time point;

(e) isolating EVs or tEVs from the sample based on detection of EV biomarkers on the EVs or tumor biomarkers on the tEVs at the second time point;

(f) determining a presence or level of a tumor drug-resistance biomarker of the EVs or the tEVs isolated at the second time point; and

(g) comparing a presence or level of the tumor drug-resistance biomarker at the first time point to a presence or level of the tumor drug-resistance biomarker at the second time point, wherein a difference in a presence or level between the first time point and the second time point indicates that the tumor’s resistance to the PARPi has changed over time.

7. The method of claim 6, wherein a presence of the tumor drug-resistance biomarker at the second time point, but not the first time point indicates that the tumor has developed a resistance to the PARPi drug over time.

8. The method of claim 6, wherein an increase in the level of the tumor drugresistance biomarker at the second time point compared to the level at the first time point indicates that the tumor’s resistance to the PARPi drug has increased over time.

9. The method of claim 6, wherein the first time point is before treatment, and the second time point is during treatment or after treatment.

10. The method of claim 6, wherein the first time point is after treatment has begun, and the second time point is during ongoing treatment.

11. The method of any one of claims 1 to 10, wherein the sample comprises subject- derived organoids or a blood sample from the subject.

12. The method of any one of claims 1 to 11, wherein the EVs or tEVs are captured on a nanosensor chip for processing.

13. The method of any one of claims 1 to 11, wherein the EVs or tEVs are labeled with antibodies that bind specifically to one or more tumor drug-resistance biomarkers.

14. The method of any one of claims 6 to 13, wherein a change in one or more tumor drug-resistance biomarkers is tracked over the course of the subject’s treatment.

15. The method of any one of claims 1 to 14, wherein the PARP inhibitor is selected from the group consisting of olaparib, rucaparib, niraparib, and talazoparib.

16. The method of claim 15, wherein an additional anti-cancer drug, e.g., a chemotherapeutic drug is added to the sample or is administered to the subject before a sample is obtained from the subject.