Liquid biopsy in ovarian cancer: recent advances in circulating extracellular vesicle detection for early diagnosis and monitoring progression.

Isolation and detection are two important and indivisible parts of EV studies. It would be ideal if detection could be achieved with raw materials such as blood and urine. However, this is very challenging to achieve with these complex biofluids, as the presence of proteins may cause the actual targets to be hard to detect. Therefore, many isolation methods involve both purification and enrichment, which make the EV concentration higher for better detection. While most studies involve a pre-isolation step before the actual analysis, there have recently been some attempts to combine isolation and analysis into one system, especially with lab-on-chip devices 62 , 63 . Physical isolation techniques are used to isolate EVs based on their physical properties like density, surface charge, or size. Conventional bulk methods based on physical isolation include ultracentrifugation, ultrafiltration, and size exclusive chromatography (SEC). Ultracentrifugation is considered the gold standard; however, it is time-consuming and has always been associated with additional issues, such as low recovery and low purity 64 . Recently, new separation technologies have been developed, mostly based on microfluidic platforms utilizing the physical properties of EVs. These new technologies include acoustic, membrane filtration, viscoelastic flow, nanowire trapping, and lateral displacement systems 35 , 65 . Whereas physical separation techniques yield higher numbers of EVs without the need for labelling or modification, they usually co-isolate different types of EVs, protein aggregates, and other particle contaminants. For example, protein bound complexes co-exist with EVs when isolated using the polyethylene glycolebased precipitation method 66 . In addition, direct isolation of cell- or tissue-specific exosomes is not possible when using physical separation methods, as they do not target surface biomarkers. Unlike the physical isolation techniques, it has been demonstrated that biological- or affinity-based separation techniques are better at isolating specific subtypes of exosomes by targeting surface proteins mainly from the tetraspanin family (e.g., CD9, CD63, and CD81) 67 . These methods are able to directly characterize the captured exosomes or lyse the exosomes for downstream analysis. However, as it is difficult to remove EVs from the binding molecules, these isolated EVs cannot be used for the functional analysis of intact EVs. Magnetic bead kits are commercially available for biomarker-specific exosome isolation (e.g. beads from ThermoFisher Scientific and System Biosciences). However, these approaches are typically expensive and require multiple steps for washing and enrichment. Recently, microfluidic devices, which bring magnetic beads into lab-on-chip systems, have been developed. These lab-on-chip systems combine all necessary steps into one device: sample loading, mixing, incubation, washing, and downstream analysis for proteins and RNAs. The lab-on-chip systems make the clinical translation of EV analysis possible 57 , 68 . Analysis of isolated exosomes is typically based on conventional detection approaches to measure the expression of exosomal proteins, such as western blot, enzyme-linked immunosorbent assays (ELISA), and flow cytometry (FCM), 69 . In ELISA, exosomes are immobilized onto a solid phase, followed by labelling with fluorescent- or enzyme-conjugated antibodies (Abs) for optical detection. In FCM, exosomes are bound to Ab-conjugated microbeads and then analyzed by measuring fluorescence of fluorescence-conjugated Abs. With these types of labelling methods, the detection signals such as absorbance (OD) or fluorescence intensity provide only the relative quantity of exosomes. Because of the small size of EVs, most FCM-based analyses still rely on microbeads to capture EVs. Microbeads enable the analysis of EVs based on biomarkers on their surface. However, existing FCM methods have limited sensitivity and resolution to analyze EVs directly, as it tends to miss or underestimate small vesicles (< 200 nm) due to “Swarm Theory” 70 . Recently, highly sensitive FCMs are under development to distinguish particles as small as 100 nm 71 so that single EVs can be interrogated. Moreover, imaging-based technology has been developed to analyze single EVs in a multiplexed format 72 . Kibria et al. developed a microFCM platform that is capable of assessing the expression of CD47 in single circulating exosomes from breast cancer patients 73 . These new technologies provide opportunities for profiling single exosome and thus, differentiate different exosome subsets. Physical analysis has been achieved for EVs as well. For physical analysis, pre-isolation to obtain a high purity EV population is particularly important. Particle size distribution and concentration are usually measured by nanoparticle tracking analysis (NTA), FCM, and tunable resistive pulse sensing 74 . NTA is a standard method for characterization and measurement of the concentration of exosomes or vesicles (< 200 nm). In NTA, a light beam illuminates the particles in the solution and the path of each particle is captured to determine its velocity and diffusivity, which will then be used to calculate the particle concentration and size distribution 75 , 76 . NTA is a simple and quick analysis. However, the results regarding size and concentration are affected by different parameters during video capture and analysis, such as camera level and threshold. In addition, the linear range for NTA to provide an accurate measurement is around 10 8 -10 9 particle/mL, which limit its application in measuring samples with low particle concentration. An alternative to NTA, tunable resistive pulse sensing (TRPS), is based on the ionic current change when a particle passes through a size-tunable nanopore. As TRPS measures individual particles, it has less strict requirements on the particle concentration. However, the particle size range that can be measured by TRPS is limited by the size of the nanopore. The nanopore may need to be changed when measuring particles in different size ranges. In addition, TRPS is not suitable for analyzing heterogeneous samples, such as plasma, as the nanopores tend to get clogged with large particles. These two techniques have recently been compared for EVs in clinical cerebrospinal fluids, suggesting that both methods are capable of assessing EVs derived from body fluids and that a multi-platform quantitation will be required to guide clinical studies 77 . Apart from multi-platform quantitation, the addition of pre-isolation procedures of exosomes such as ultracentrifugation and SEC, or precipitation reagents (such as polyethylene glycol) have also been suggested to better assess the size and distribution of EVs. Nevertheless, it has been shown that neither the total number of EVs nor the size of EVs is accurate in differentiating different status of cancers and healthy controls 43 . Thus, these physical parameters need to be combined with molecular information for clinical relevance. Overall, bulk methods based on counting or labelling have limitations such as being time-consuming, labor-intensive, or insensitive. These limitations are greatly hindering the translation of current exosome analytical methods into clinical settings where real-time monitoring and high-throughput analysis is required for samples with low exosome abundance.

Surgical tissue biopsies are invasive procedures and can be associated with complications such as bleeding and infection 28 . In addition, biopsies are often difficult to perform on organs that lie deep within the body and the use is limited as they can give false negative results due to sampling bias 29 . Compared to conventional tissue biopsy, liquid biopsy is growing in popularity because it is minimally invasive, easy to use, and can have high throughput. ctDNA in the plasma of OC patients can identify relapse or drug resistance well before clinical symptoms appear, enabling earlier intervention and better patient outcomes 30 . Liquid biopsies measure various tumor biomarkers such as proteins, nucleic acids, cells, and EVs in body fluid like blood. Thus, liquid biopsies are advantageous over traditional tissue biopsies as blood samples can be easily collected longitudinally and in large quantities, making it an attractive platform for large-scale screening of tumor-specific mutations 31 , 32 . It also has the potential of providing new insights into prognosis, patient follow-up, treatment response, and more recently, early diagnosis and population screening 33 . Liquid biopsies of CTCs, ctDNA, and EVs are promising for early-stage cancer detection and real-time monitoring the dynamics of cancer progression and metastasis 23 , 34 . It has been shown that cancer cells release EVs containing cancer-specific contents that can be easily isolated from various body fluids 35 . However, using either CTCs or ctDNA as cancer biomarkers faces multiple technical and translational challenges. First, scarcity and heterogeneity of CTCs make the isolation and characterization of CTCs extremely hard 36 . Second, high fragmentation, low abundance, and low stability of ctDNA largely hampered the utility of ctDNA in routine clinical practice 37 , 38 . Compared with CTCs and ctDNA, EVs possess advantages in terms of abundance, stability, and accessibility. First, EVs are abundant (10 8-13 exosomes/mL) in plasma and other body fluids. Secondly, EVs are very stable 39 and can be stored at -80C° for months and even years while maintaining protein and nucleic acid quality. Furthermore, the contents of EVs are tumor-specific and correlate with tumor staging and prognosis 40 . In addition, EVs are broadly distributed in body fluids and thus, can be easily obtained. From the same type of tissue, cancer cells were found to shed more EVs compared to normal cells, indicating EVs are a much more abundant biomarker source in liquid biopsy compared to CTCs 41 . Therefore, based on these merits, more emphasis has been put on EVs as a biomarker source for liquid biopsy of cancer in recent years 42 , 43 .

Various new technologies have recently been developed to improve the sensitivity and throughput for EV analysis, such as microfluidic technology, which has previously been shown to have unique advantages in exosome separation, genomic and proteomic analysis, as well as quantitative biology. It also features low sample volume requirement and simple sample processing, which makes it feasible for point-of-care clinical utilities. The following approaches have been recently developed for OC exosome characterization and shown promise in the clinical setting for OC diagnosis and prognosis. The nPLEX assay is a label-free, high-throughput approach for quantitative analysis of exosomes 58 . This method is based on transmission surface plasmon resonance to detect proteins on the surface or in the lysates of exosomes. This approach had improved sensitivity compared with conventional modalities and could be portably operated when integrated with miniaturized optics. Im et al. demonstrated that nPLEX could identify OC-derived exosomes from ascites in patients by detecting CD24 and EpCAM, suggesting its potential for diagnostics ( Figure 1 ). Compared to conventional methods, the nPLEX technology has advantages such as high sensitivity, label-free exosome analyses, and continuous real-time monitoring of molecular markers. With iMEX assay exosomes were immunomagnetically captured from OC patient samples and assessed through an electrochemical reaction. Combining immunomagnetic enrichment and enzymatic amplification, the approach demonstrates high sensitivity, cell-specific detection, sensor miniaturization, and high-throughput ability for exosome measurements 59 . The iMEX is a portable exosome detection system with the capacity to perform measurements in parallel. The sensor can simultaneously detect multiple protein markers within an hour while consuming only 10 μL of plasma per marker, which outperforms conventional methods in terms of sensitivity and speed. This group found higher levels of EpCAM and CD24 in EVs from OC patients than those from healthy controls, and both metrics showed high correlation ( Figure 2 ). In addition, they also examined iMEX's potential for real-time monitoring EV markers EpCAM and CD24 in plasma of OC patients before and after drug treatment. Their results suggested that the “nonresponding” patients had high expression levels of EpCAM and CD24 compared with the “responding” patients ( Figure 2 C ). Compared with nPLEX, iMEX has lower sensitivity and throughput, but is less complex and does not require nanofabrication, which makes it an affordable and miniaturized platform for on-site exosome detection. ExoSearch is a simple microfluidic approach for the rapid preparation of blood plasma exosomes for in situ , multiplexed detection using immunomagnetic beads 57 . ExoSearch chip has been employed for plasma-based diagnosis of OC by multiplexed evaluation of the expression levels of CA-125, EpCAM, and CD24 on the surface of exosomes in 20 OC patients, which demonstrated superior diagnostic power (AUC = 1.0, p = 0.001). The ExoSearch chip has the capability to perform simultaneous and quantitative evaluation of a biomarker panel from the same exosome subpopulation with improved reproducibility. In addition, this assay can acquire different subpopulations of exosomes from a wide range of input volumes (10 μL to 10 mL), largely facilitating the downstream molecular analysis and profiling. However, given the small number of patients recruited in the study, future studies with a large-scale cohort is required to further validate the diagnostic value of the ExoSearch chip. Later, this group developed another sensitive microfluidic platform based on a new graphene oxide/polydopamine (GO/PDA) nano-interface (nano-IMEX), which could discriminate OC patients from healthy controls by using 2 μL plasma without sample processing 60 . This suggests that this platform could provide a more robust assay to evaluate exosomes for non-invasive detection and precision treatment of OC. Kabe Y et al. recently designed a novel device, the ExoCounter, to quantify the number of exosomes in the sera of OC patients. In this system, exosomes can be captured in the groove of an Ab-coated optical disc, labeled with Ab-conjugated magnetic nanobeads, and then counted with an optical disc drive 78 . This team demonstrated that this new approach could detect specific exosomes derived from cell supernatants or human serum without any enrichment procedures. In addition, ExoCounter had high detection sensitivity and linearity compared with conventional detection methods such as ELISA or FCM. Using ExoCounter, the CD9/HER2-positive exosomes were shown to be significantly increased in patients with OC compared with healthy controls and noncancer disease patients. Therefore, this method is very suitable for liquid biopsies of OC exosome biomarkers for diagnosis and progression monitoring. A herringbone-grooved microfluidic device has recently been developed for direct isolation of exosomes using biomarkers CD9 and EpCAM from small volumes of serum of high-grade serous ovarian cancer (HGSOC) patients. Using this device, they found that both total and EpCAM + exosome numbers increase concurrently with disease progression in HGSOC 79 . This approach can be used to isolate intact and label-free biomarker specific exosomes for predicting HGSOC disease stages, as well as facilitating downstream functional studies. In comparison with traditional isolation methods, this platform features a rapid (< 20 minutes for capture and release) and cost-effective method with a high yield and specificity and low sample volume requirement (< 100 μL) to distinguish significant differences in HGSOC disease stages, making itself suitable for clinical applications. In addition, as the exosomes captured by the platform are intact and label-free, this method allows further downstream characterization and experimentation, both on and off chip. In summary, all new techniques recently developed for EV detection hold promise for OC early diagnosis and monitoring cancer progression. However, new OC exosomal markers should continue to be tested using these technologies and a large number of OC samples need to be used for validation studies to confirm their clinical significance.

While exosomes show great promise as biomarkers for OC diagnosis and real-time progression monitoring, there are still several limitations that need to be overcome prior to more widespread clinical application. First, standardized and consistent methods need to be established for the isolation and enrichment of tumor-derived exosomes from blood samples, as no clear consensus has been reached regarding the optimal method for isolation and quantification of exosomes 11 . Second, the identified existing exosome protein biomarkers in OC need to be validated in a large set of sample cohorts to find the impact on clinical outcomes such as improved early detection, progression-free survival, or overall survival rates. Some of the new technologies mentioned in the previous sections have follow-up studies that involve clinical sample cohorts. For instance, the nPLEX assay developed by Im et al. 58 has been applied to pancreatic ductal adenocarcinoma in more than 100 clinical samples 80 . Other new technologies need to be validated in clinical studies. Third, the time and costs for exosome processing and analysis should be significantly reduced for clinical application and these non-invasive detection methods should be accurate and fast 42 . Some of the new technologies described above have effectively addressed this challenge, aiming towards developing rapid and cost-effective tests. In addition, many other factors such as stress, hypoxia, tumor types, and growth patterns can influence the secretion of exosomes and should be taken into consideration during processing and analysis 81 , 82 . Despite these challenges, exosomes have shown significant potential as future liquid biopsy biomarkers for OC and further research and development is warranted in this area.

Ovarian cancer (OC) is one of the most lethal gynecological malignancies. It is the fifth leading cause of cancer-related deaths among females, affecting over 290,000 women worldwide annually 1 with an estimated 22,240 new cases and 14,070 deaths in the United States in 2018 2 . Due to the lack of early symptoms, physical signs, and effective screening approaches for early diagnosis, approximately 70% of OC cases are not diagnosed until they are in advanced stages, which only have a 20% 5-year survival rate. However, if diagnosed at early stages, the 5-year survival rate for Stage I and II OC is 89% and 71%, respectively 2 . Therefore, early diagnosis using effective biomarkers and screening approaches is of high importance and may improve the prognosis of a large number of OC patients. The diagnosis of OC is mainly based on levels of biomarker CA-125 in blood and imaging 3 , 4 . CA-125, also known as MUC16, is the most clinically utilized biomarker for monitoring the response to treatment and detecting disease recurrence in OC 5 . However, CA-125 levels are not always increased in the early stages of OC and not every OC patient shows elevated CA-125 levels. In addition, some other diseases such as endometriosis, inflammation, and other types of cancers 6 - 8 can also cause elevated CA-125 levels. Furthermore, even with using CA-125 as a screening marker, the overall survival rate in OC has not significantly changed in clinical trials 4 , 9 . As a result, no professional group recommends screening ovarian cancer using CA-125 in the general population. Therefore, it is of great importance to find new approaches to detect early stage OC. Extracellular vesicles (EVs) including exosomes, microvesicles, and other membranous structures are abundantly released into the extracellular space by almost all types of cells. EVs carry complex biological information from their original cells and are useful sources for cancer diagnosis in a non-invasive manner 10 . According to the International Society of Extracellular Vesicles (ISEV), the term “extracellular vesicles” is the appropriate terminology for heterogeneous populations of vesicles isolated from cell culture supernatants or physiological fluids 11 . Throughout this review, exosomes will be referred to as EVs. Exosomes are cell-secreted membranous nanoscale vesicles with diameters of 50-150 nm that contain mRNA, microRNA, small interfering RNA, and proteins 12 - 15 . These exosomal contents are representative of its originating cell and contribute to intercellular communications 16 . Exosomes attract considerable interest in the research community due to their role in regulating multiple physiological processes and mediating systemic dissemination in various cancers 17 . Several reports have demonstrated that exosomes exist in blood and ascites of OC patients 18 , 19 . In addition, exosomes and exosomal cargoes, such as microRNAs, were found to play a crucial role in disease progression and potentially facilitate chemoresistance in OC 20 - 22 . Therefore, OC-derived exosomes have the potential to be used as biomarkers for the early detection of cancer and follow-up monitoring. Liquid biopsy, a recent and hot topic in cancer detection, has been considered for the early diagnosis of cancer 23 . Generally speaking, liquid biopsy involves the collection and analysis of circulating tumor cells (CTCs), circulating tumor DNA (ctDNA), circulating cell-free microRNAs (cfmiRNAs), and exosomes 24 . Liquid biopsy has already been used in OC research 24 . Both CTCs and ctDNAs in OC have been intensively studied for clinical significance in the last two decades and the advances in the field have been recently reviewed 25 - 27 . This review highlights the recent progress in new techniques for OC EV detection and mainly focuses on EV protein biomarkers for OC early detection, monitoring cancer progression, and personalized therapy.

Liquid biopsy is nearly ready to offer a robust, yet minimally invasive tool for the diagnosis and comprehensive management of OC. Apart from being able to provide valuable information for diagnosis when the tumor is less accessible, blood-based exosome tests may also allow for real-time monitoring of the tumor evolution and evaluation of treatment efficacy. Although most of the biomarkers available today require prospective validation, the development of non-invasive EV-based liquid biopsy has already emerged and paved the way to improve the early detection, evaluation of response to therapy, prognosis, and outcome in OC patients. As surrogates of cancer cells, exosomes are promising for precise and personalized cancer diagnosis and real-time monitoring cancer progression. Using a panel of identified exosomal protein markers as a “cancer signature” may provide improved detection in screening OC for early diagnosis. The EV cargo provides a promising source for the discovery of liquid biopsy biomarkers. The rapid advances in next-generation “omics” and EV capture platforms are the driving approaches for disease stratification, diagnosis, and monitoring. Further advancements in EV isolation methods that potentially prevent overestimation and contamination of EVs may allow the study of discrete EVs from body fluids, hence holding great promise for future diagnostic applications, where isolation and examination of individual EVs are paramount. More rapid and defined EV isolation procedures have been recently developed. This should enable the seamless integration of EV isolation and analyses into clinical diagnostic pipelines. This is crucial since time-consuming isolation procedures that require expensive, specialized equipment (i.e. ultracentrifugation) are unlikely to be feasible for routine clinical practice. In addition, specific exosomal cargo molecules (proteins, RNAs, lipids, etc.) are likely to be identified/validated in the context of defined clinical questions.