Single extracellular vesicle imaging via rolling circle amplification–expansion microscopy

Wu, Jiacheng; Dou, Quanhao; Mao, Miao; Wan, Xin; Wu, Minhao; Hu, Tony Y.; Zhang, Yuanqing

doi:10.1038/s41467-025-62613-0

Download PDF

Article
Open access
Published: 13 August 2025

Single extracellular vesicle imaging via rolling circle amplification–expansion microscopy

Nature Communications volume 16, Article number: 7498 (2025) Cite this article

9521 Accesses
26 Altmetric
Metrics details

Subjects

Abstract

Extracellular vesicles (EVs) from biological fluids can provide critical information for minimally invasive diagnostics and treatment monitoring, but their nanoscale size, low biomarker abundance, and heterogeneity pose challenges. Here, we integrate rolling circle amplification with expansion microscopy (RCA–ExM) to achieve super-resolution multi-omics profiling of single EVs using conventional fluorescence microscopy. Sensitive multimodal biomarker detection is achieved by employing RCA to detect switch hairpin probe-labeled EV membrane proteins, and EV-liposome fusion to detect EV miRNAs via delivery of specific molecular beacons and a signal-amplifying enzyme circuit. Next, hydrogel-mediated expansion is employed to enlarge the fused EVs to permit single-EV detections. RCA–ExM quantitation of miRNA-21 levels in EpCAM⁺ PD-L1⁺ plasma EVs from a clinical cohort (n = 86) successfully distinguishes cancer patients from healthy donors and differentiates 3 categories of immunotherapy efficacy. RCA–ExM therefore exhibits significant promise for more sensitive and specific diagnostics, and treatment monitoring applications.

Extracellular vesicles as tools and targets in therapy for diseases

Article Open access 05 February 2024

Using single-vesicle technologies to unravel the heterogeneity of extracellular vesicles

Article 16 June 2021

Multiplex protein profiling method for extracellular vesicle protein detection

Article Open access 14 June 2021

Introduction

Extracellular vesicles (EVs) are membrane-bound structures secreted into the extracellular milieu by all cells^1,2. EVs carry bioactive molecules, including proteins, lipids, and nucleic acids that play essential roles in intercellular communication and provide real-time insight into the physiological and pathological states of their parental cells³. Notably, EVs are reported to play important roles in the initiation, development, metastasis, and drug resistance of various tumors, including breast cancer, prostate cancer, lung cancer, liver cancer, and colorectal cancer^3,4,5,6,7. EVs are thus emerging as a promising source of liquid biopsy–based biomarkers for early cancer diagnosis, prognosis, and treatment monitoring assays. However, new approaches that can profile distinct EV biomolecule types are needed to advance cancer research and develop better EV-based clinical applications.

EVs exhibit significant compositional heterogeneity, particularly in their surface proteins and encapsulated microRNAs (miRNAs or miRs) that reflect their cells of origin and tumor-derived EVs contain biomolecules that serve as promising biomarkers for cancer diagnosis and therapy. For example, programmed death ligand 1 (PD-L1) protein expression, which is recognized as both an effective cancer diagnostic indicator and as a reliable predictor of an immunotherapy response⁸, is expressed on tumor-derived EVs. However, PD-L1 has limited specificity for tumor-derived EVs, since it is also expressed on EVs secreted by immune cells– and mesenchymal stem cells⁹, and as a soluble protein. EVs secreted by various cancers (e.g., liver cancer, esophageal squamous cell carcinoma, glioma, and breast cancer) are also enriched in various miRNAs, which could serve as diagnostic markers^{10,11,12,13,14}. Notably, miR-21 levels of PD-L1 positive EVs can distinguish cancer patients from healthy donors and correlate strongly with tumor burden and immunotherapy efficacy¹⁰. Hence, the dissection of EV biomarker heterogeneity can have substantial clinical utility for disease diagnosis and precision medicine.

Despite the potential utility of such EV biomarkers, two issues pose significant challenges for the development of sensitive cancer diagnostic EV applications. First, tumor-derived EVs can be rare and express biomarkers at relatively low levels, necessitating the use of highly sensitive assays. The heterogeneity of EV populations in biological specimens^15,16,17,18 can also lead to substantial variation or inaccuracy when these samples are evaluated using conventional bulk analysis methods, such as western blotting, nucleic acid amplification, and sequencing, which analyze biomarker expression across the entire EV sample^19,20,21. New technological approaches that can analyze the heterogeneity of EV subpopulations have been developed to address this problem^22,23,24, but their sensitivity and specificity are inferior to single-EV analysis methods, and can fail to provide precise and detailed information about specific EV subtypes. However, single-EV phenotypic analysis methods, such as nano-flow cytometry, can also struggle with sensitivity and resolution when used to detect small changes in EV biomarker levels. Second, the diminutive size of EVs, which typically range from 30 to 200 nm, poses additional challenges for single-EV detection approaches, including the detection and resolution of biomarker signal from single EVs^25,26,27. Fluorescence microscopy, which is widely employed as a fundamental technique in biological research^28,29,30,31, is constrained by its diffraction limit, preventing its use for direct detection and evaluation of single EV biomarker signals. Super-resolution microscopy can achieve image resolutions to 30–50 nm, but its high cost, specific hardware requirements, and sophisticated data processing procedures have restricted its widespread use^32,33,34,35.

In this work, we develop a method that employ rolling circle amplification (RCA) and EV-liposome fusion-mediated delivery of molecular beacons and a signal-amplifying enzyme circuit to sensitively detect surface biomarker and lumenal miRNA signal via an expansion microscopy (ExM)³⁶ method that enables for super-resolution EV imagery with a conventional fluorescence microscope. The RCA–ExM method facilitates sensitive and specific detection of EVs with distinct surface biomarker and miRNA phenotypes, enabling evaluation of heterogeneous EV populations at the single-EV level. We show that RCA–ExM analysis reveals strong cancer diagnostic performance and immunotherapy efficacy prediction by miR-21 expression in EpCAM⁺ PD-L1⁺ plasma EVs. This approach can thus both open new avenues for EV research and enhance the development and clinical translation of EV-based liquid biopsy diagnostic applications.

Results

RCA–ExM assay design

Our proposed RCA–ExM assay was designed to require only three streamlined procedures to detect of multiple biomarkers on single EVs (Fig. 1A). First, target EV were captured from plasma, culture supernatant, or some other specimen type via an allosteric aptamer targeting a tumor-associated EV surface factor (e.g., EpCAM). A target-specific switch hairpin probe (SP) was first hybridized to a second tumor-associated factor (e.g., PD-L1) on the surface of these captured EVs to provide a substrate for subsequent RCA amplification. These SP-labeled EVs were then hybridized with liposomes containing molecular beacons (MBs) and dual-specificity nucleases (DSNs) to generate a signal-amplifying enzyme circuit in the resulting Lipo-EV fusion vesicles to permit sensitive in situ detection of target miRNAs in the EV lumen. This step allowed precise quantification of tumor-derived EVs and avoided variations arising from miRNA degradation by ribonucleases (RNases) and interference from soluble miRNA in the specimen. RCA was then initiated on the SPs bound to the surface of the Lipo-EVs to generate multiple binding sites for a reporter modified with acrydite and fluorophore groups. Finally, these fluorescently labeled RCA-Lipo-EVs were released from the chip through Nb.BbvCI digestion, incorporated into a hydrogel matrix via their acrydite groups and subjected to an expansion process to allow the resolution of enlarged single RCA-Lipo-EVs by conventional fluorescence microscopy. This method offered significant advantages over existing biosensing assays for EV biomolecules by providing superior analysis of single-EV compositional heterogeneity while reducing reliance on sophisticated instrumentation and complex data processing methods, and its utility was validated by in studies designed to profile multi-phenotypic EVs, and diagnose cancer and evaluate immunotherapy efficacy (Fig. 1B).

**Fig. 1: Schematic of the RCA–ExM method for in situ single-EV biomarker profiling.**

Capture and membrane protein labeling of tumor-derived EVs

To detect tumor-derived PD-L1⁺ EVs in complex biological fluids, we developed a dual-labeling approach in which a nanocomplex formed by a thiol-modified single-stranded DNA (SH-ssDNA) and a partially complementary EpCAM-specific aptamer served as a tumor-specific marker, while a PD-L1 specific switch hairpin probe (SP) containing three domains (aptamer, spacer (S), and linker (L) domains) served as marker of immune function. In its non-active state, this PD-L1-specific SP produces a hairpin structure that blocks access to its L domain. However, upon target recognition, this SP undergoes an allosteric transformation that exposes its L domain to provide a substrate for RCA, which in turn creates multiple binding sites for the reporter probe. This SP-mediated RCA process thus allows rapid and selective labeling of EVs for subsequent analyses.

To streamline EV capture and labeling, and facilitate in-situ detection, we designed and fabricated a detachable polydimethylsiloxane (PDMS) chip that was modified with EpCAM-specific aptamer/SH-ssDNA nanocomplexes containing Nb.BbvCI cleavage sites to permit direct capture, and non-destructive release of tumor-derived EVs from biological specimens (Fig. 2A and Supplementary Fig. 1). EpCAM⁺ EVs captured on this chip were then incubated with the PD-L1-specific SP, RCA reagents, and reporter to allow sensitive detection of EpCAM⁺ EVs. After releasing the captured RCA-Lipo-EVs from the chip, the chip wells served to confine the added hydrogel mixture during its polymerization around these vesicles, after which the PDMS layer was removed to allow lateral expansion of the hydrogel on the glass substrate. This integrated approach eliminates the necessity for disruptive EV transfer steps while preserving spatial registration throughout the workflow.

**Fig. 2: Capture and surface labeling of tumor-derived EVs.**

Successful assembly of the EpCAM-specific aptamer/SH-ssDNA nanocomplex and its Nb.BbvCI-mediated specific cleavage was confirmed by agarose gel electrophoresis analysis (Supplementary Fig. 2). Fluorescence microscopy was employed to further analyze the capture behavior and release efficiency of FAM-labeled nanocomplexes (Fig. 2B and Supplementary Fig. 3). Based on these results we conjugated all future PDMS chips with 1 μM aliquots of the EpCAM-specific aptamer/SH-ssDNA nanocomplex and incubated with Nb.BbvCI to achieve 85% release efficiency.

We next evaluated the ability of the PD-L1-specific SP to label PD-L1⁺ EVs, using a RCA strategy to increase detection sensitivity given the low protein abundance on EVs. We first demonstrated that this SP underwent a target-induced conformational change to initiate a toehold-mediated RCA reaction (Supplementary Fig. 4). Flow cytometry and fluorescence microscopy data revealed this RCA-based detection approach yielded significantly stronger signal than conventional detection methods that did not employ RCA (Supplementary Fig. 5). Fluorescence microscopy images also clearly detected RCA-based labeling of the membrane of PD-L1⁺ A375 cells, but not PD-L1⁻ MCF-7 cells (Pearson correlation coefficient > 0.5) (Fig. 2C) to support the sensitivity and specificity of this PD-L1 labeling strategy. However, subsequent optimization studies found that reducing the length of the fluorescent reporter that binds the RCA sequence could further increase target sensitivity (Supplementary Fig. 6), to yield an RCA approach that enhanced target signal 4.3-fold more than signal detected from control samples that did not employ RCA (Supplementary Fig. 7). This shortened reporter was therefore used in all subsequent EV detection studies.

EVs derived from two cancer cell lines with distinct PD-L1 and EpCAM expression profiles (A375 [+/+] and MCF-7 [−/+]) were evaluated for their potential to serve as positive and negative controls for EV PD-L1 expression. Flow cytometry analyses performed with a FAM-labeled PD-L1-specific aptamer and a Cy5-labeled EpCAM-specific aptamer detected PD-L1 signal only on the A375 cell EVs, EpCAM signal on both the A375 and MCF-7 cell EVs, and neither PD-L1 or EpCAM signal on the Jurkat EVs employed as the negative control (Fig. 2D and Supplementary Fig. 8), consistent with PD-L1 and EpCAM expression detected on these EVs in previous reports^37,38. EVs isolated from A375 and MCF-7 cell cultures also revealed overlapping size distributions (Supplementary Fig. 9) and zeta potential characteristics (Supplementary Fig. 10), indicating they lacked physical and electrical potential differences that could confound the results of studies employing these EVs as positive and negative controls for EV PD-LI expression.

We next employed a nanocomplex-modified PDMS chip to capture MCF-7 and A375 cell EVs, and then in situ labeled these EVs using our PD-L1-specific SP/RCA strategy, and the lipid-selective fluorescent dye DiI to examine the specificity of the detected PD-L1 signal, and its decrease after a Nb.BbvCI-mediated nanocomplex cleavage reaction. PD-L1 signal closely aligned with EV membrane DiI signal on A375 but not MCF-7 cell EVs, consistent with the differential PD-L1 expression of these two EV samples. Further, both PD-L1- and EV membrane-associated signal decreased in to near background levels when A375 EV loaded chips were incubated with Nb.BbvCI, consistent with efficient release of the dual labeled EVs (Fig. 2E). Taken together, these findings indicate that our dual-labeling approach, which employs a EpCAM-specific nanocomplex to selectively capture and release tumor-derived EVs and SP/RCA to label captured EVs that express both EpCAM and PD-L1, permits efficient EV capture, in situ labeling, and release.

EV miRNA detection by liposome transfer of an enzyme circuit signal amplification assay

EV miRNAs, which can also serve as cancer biomarkers, are typically analyzed in bulk EV miRNA preparations that are subject to degradation and contamination during the isolation process. Such isolates also cannot detect miRNA differences associated with specific EV subtypes, which can mask changes in less abundant but biologically important EV populations. We therefore employed a liposome-EV strategy to deliver an enzyme circuit–based signal amplification assay to captured EVs to permit in situ detection of EV miRNA signal and its direct alignment with the corresponding SP/RCA signal on matching EVs. In this assay design, reagent loaded liposomes are induced to fuse with and transfer their contents to captured tumor-specific EVs, allowing a transferred MB to bind its complementary EV miRNA target, which allows a transferred DSN to cleave the DNA strand of this DNA-RNA heteroduplex^39,40, and releasing and dequenching the reporter fluorophore and allowing the previously bound miRNA to bind another reporter molecule and repeat this cycle, yielding a concentration-dependent signal amplification (Fig. 3A and Supplementary Fig. 11).

**Fig. 3: Liposomal transfer of signal amplification circuit for in situ miRNA detection.**

Characterization of the preparation of cationic liposomes and analyzing their fusion with EVs to determine their size distribution and zeta potential, revealed that the EVs and liposomes were of similar size, while the vesicles detected after fusion were larger size and had zeta potentials that were intermediate to the original vesicles (Fig. 3B, C). Furthermore, MB-loaded liposomes displayed lower zeta potential magnitudes than empty liposomes, confirming both successful preparation of the liposomes and effective fusion with EVs (Fig. 3C). Transmission electron microscopy (TEM) analysis revealed that the liposome and EV samples exhibited characteristic spherical and cup-shaped morphologies, and that vesicle size notably increased following the fusion reaction (Fig. 3D). Electrostatic interactions between these liposomes and EVs should favor liposome-EV fusion rather than autologous EV-EV or liposome-liposome fusion events EV, as suggested by our zeta potential data. However, to further confirm that liposome-EV fusions occur, we next used a fluorescence resonance energy transfer (FRET) approach to detect the dilution of the EV membrane by liposome membrane material during the liposome-EV fusion reaction. In this approach, EVs were labeled with the FRET donor and acceptor dyes DiO and DiI so that, prior to liposome-EV fusion, the signal produced upon DiO excitation (510 nm) was efficiently transferred to DiI to stimulate a higher wavelength signal (563 nm), whereas a membrane dilution effect produced upon liposome-EV fusion increased the distance between these dyes to decrease this DiO to DiI signal transfer and increase DiO signal while decreasing DiI signal (Fig. 3E). Our results indicate that DiO signal increased with time after the initiation of the fusion reaction, consistent with Lipo-EV fusion events.

Specific detection of our EV miRNA target requires our assay’s DSN to cleave the DNA strand of the resulting reporter DNA-microRNA heteroduplex, but the dsDNA region of the MB stem-loop structure could also serve as DSN target to yield high background or false-positive signal in the absence of its miRNA target. Since previous studies have reported that dsDNA target < 8 base pairs in length are resistant to DSN digestion⁴¹, we selected a MB (MB-21) with a 5 base pair stem region for this study. Our subsequent analysis confirmed that MB-21 was only cleaved by our assay DSN after it formed a heteroduplex with its miRNA target (Supplementary Fig. 12).

To evaluate the signal amplification performance of the MB/DSN circuit in situ, we incubated miR-21⁺ EVs with MBs, MB-loaded liposomes, and MB/DSN-loaded liposomes (Fig. 3F). This analysis revealed that miR-21⁺ EVs incubated with MB/DSN-loaded liposomes produced ~8.3- and ~5.5-fold more signal, respectively, than those incubated with free MB or MB-loaded liposomes, indicating that DSN can significantly increase the detected EV miR-21 signal intensity. DSN-mediated miR-21 signal amplification was even more apparent when the captured EVs were incubated with MB, or MB- or MB/DSN-loaded liposomes and analyzed by fluorescence microscopy (Fig. 3G). This analysis detected substantial miR-21 signal and miR-21 and PD-L1 signal colocalization in the MB/DSN-liposome group, faint miR-21 signal and poor colocalization in the MB-liposome group, and no apparent miR-21 signal in the free MB group. Quantitative analysis confirmed that our strategy combining signal amplification via the DSN circuit with membrane fusion enables highly sensitive detection of miR-21 in tumor-derived PD-L1–positive EVs (Fig. 3H).

RCA–ExM for super-resolution imaging of single EVs

Both the liposome fusion and RCA labeling strategies somewhat enlarged the effective size of the modified EVs, these increases were not sufficient to achieve adequate resolution for imagery with a conventional microscope. We therefore integrated our RCA-Lipo-EV target detection approach with an ExM procedure to enlarge these labeled vesicles to the extent that they could be resolved as single particles. In this modified approach, EVs would be captured, labeled by EV fusion and RCA, released, hydrogel embedded, and then subjected to a hydrogel expansion procedure in an attempt to generate a streamlined and reproducible EV labeling and expansion procedure that permits super-resolution imaging of individual EVs. To test this approach, RCA-Lipo-EVs were first anchored to hydrogel polymer matrix via acrydite groups incorporated into assay reporter molecules bound to repetitive sites in the PD-L1-bound RCA products (Fig. 4A). This vesicle-embedded hydrogel matrix was then subjected to expansion to induce uniform expansion of the bound RCA-Lipo-EVs to permit super-resolution imaging with a conventional fluorescence microscope.

**Fig. 4: RCA–ExM super-resolution imaging and phenotypic profiling of single EVs.**

Subsequent analysis of the hydrogel expansion characteristics of unmodified and RCA-Lipo-EV- vesicle-embedded hydrogel samples during the osmotic swelling procedure, found that the vesicle-embedded hydrogels revealed slower and reduced overall expansion than unmodified hydrogels (Fig. 4B). RCA-Lipo-EVs bound to these hydrogels thus may have served as an additional cross-linking agent that partially retarded and restricted the relative motion of the hydrogel polymers during the expansion process, implying that force generated by the relative motion of these polymers was transferred to the RCA-Lipo-EVs, which could have also led to their proportional expansion. We therefore next analyzed potential morphological changes in these vesicles after the hydrogel swelling process by fluorescence microscopy to evaluate the continuity and distribution of membrane protein and lipid signal on these vesicles as a surrogate marker for EV membrane integrity. This analysis detected good colocalization of these signals on particles that demonstrated radially symmetric signal distributions consistent with intact vesicle morphologies after the hydrogel swelling procedure (Supplementary Fig. 13). Fluorescent microscope images captured for vesicle-embedded hydrogels before and after the swelling procedure also revealed no significant loss in PD-L1 signal after this procedure (Supplementary Fig. 14), further suggesting that this process did not disrupt the RCA-Lipo-EVs, which could lead to the lysis of these vesicle and the potential diffusion of the surface signal with distinct hydrogel polymers during the expansion process. Subsequent ExM analysis of the PD-L1 and miR-21 signal of A375 cell EVs by fluorescence microscopy, revealed that this procedure produced a consistent increase in their mean diameter, resulting in a clear distinction between surface PD-L1 and internal miR-21 signal localization, with the latter providing additional evidence of vesicle membrane integrity (Fig. 4C and Supplementary Fig. 15).

Multiple markers may be required to resolve EVs populations secreted by cells with distinct phenotypes, and we therefore tested the ability of a multiplex RCA-Lipo-EV assay that recognized two cancer-associated EV surface proteins (EpCAM and PD-L1) and miR-21 to accurately distinguish EVs secreted by four cancer cell lines (Supplementary Fig. 16). EVs were first captured by a nanocomplex specific for CD63, a general EV marker, and then incubated with EpCAM and PD-L1-specific SPs, and acrydite-modified fluorescent reporters that bound repetitive sites in the RCAs produced by these SPs to yield green (FAM) and blue (Atto 455) signal, respectively. EpCAM- and PD-L1-associated fluorescent signal detected in this assay accurately revolved EVs secreted by A375 (EpCAM⁻ PD-L1⁺), U251 (EpCAM⁻ PD-L1⁺), MCF-7 (EpCAM⁺ PD-L1⁻), and Jurkat (EpCAM⁻ PD-L1⁻) cells⁴² (Supplementary Fig. 17). RCA-ExM analyses performed with these three markers could resolve EVs with seven distinct phenotypes via differential expression of these biomarkers (Fig. 4D and Supplementary Fig. 18).

Cancer diagnosis and immunotherapy evaluation based on RCA–ExM method

We next employed this RCA-ExM assay approach to analyze plasma samples collected from 13 healthy donors and 73 cancer patients according to their RECIST-defined immunotherapy responses: partial response (PR, n = 17), stable disease (SD, n = 15), and progressive disease (PD, n = 41) (Supplementary Tables 1 and 2). Flow cytometry results detected a 7.95-fold upregulation of EpCAM expression on the cancer vs. control group plasma EVs, despite substantial signal overlap between these groups (P = 0.0059, Supplementary Fig. 19), consistent with results from prior reports^43,44. However, miRNA expression levels detected in the plasma EpCAM⁺ PD-L1⁺ EVs of these individuals clearly distinguished the healthy donors and the PR, SD, and PD cancer patient groups (Fig. 5A). Further, EpCAM⁺ PD-L1⁺ EV miR-21 expression revealed greater performance in resolving these groups than EpCAM⁺ EV PD-L1 expression, as there was no miR-21 signal overlap between any of these group but substantial PD-L1 signal overlaps between the PR and SD groups (Fig. 5B, C). Finally, to assess whether miR-21 expression in EpCAM⁺ PD-L1⁺ EVs could distinguish distinct immunotherapy outcomes, we constructed receiver operating characteristic (ROC) curves using the pooled data from 73 cancer patients. Results of this analysis indicated that miR-21 expression levels detected in EpCAM⁺ PD-L1⁺ EVs achieved perfect discrimination for all immunotherapy responses comparisons (AUCs = 1.00 for PR vs. SD, SD vs. PD, and PD vs. PR; Fig. 5D, left). Conversely, PD-L1 expression in EpCAM⁺ EVs performed well in classifying SD vs. PD (AUC = 0.969) and PD vs. PR (AUC = 0.941), but its discriminatory capacity was notably diminished when differentiating PR from SD within this cohort (AUC = 0.718) (Fig. 5D, right).

**Fig. 5: RCA-ExM-based cancer diagnosis and evlaution of immunotherapy response.**

Finally, we performed an RCA-ExM analysis of miR-21 and PD-L1 expression levels of EpCAM⁺ PD-L1⁺ plasma EVs from five cancer patients with longitudinal plasma samples collected before and after anti-PD-1 or anti-PD-L1 immunotherapy (Supplementary Fig. 20). This analysis revealed consistent pre- to post-therapy EV PD-L1 signal increases regardless of their post-therapy response (PR or PD) or specific therapy (anti-PD-1 or anti-PD-L1). However, only the single patient with a PR outcome exhibited an EV miR-21 expression decrease, whereas all the patients with PD responses demonstrated EV miR-21 increases.

Discussion

EVs secreted by tumor cells can carry an array bioactive molecules that reflect their phenotype, or which can be transferred to recipient cells to promote tumor growth and metastasis, and are thus highly promising biomarker targets. This has sparked significant EV research over the past decade, but EV-based diagnostic technologies still face major challenges^31,45. For example, most rapid analysis approaches, including flow cytometry, suffer from low sensitivity and limited ability to resolve distinct EV subpopulations due to the nanoscale size of these vesicles and their complex compositions. Further, most EV analysis methods do not provide detailed molecular information at single-particle resolution, hindering accurate assessment of EV heterogeneity. There is thus an urgent need for more advanced and sensitive detection methods that can overcome current technological constraints to harness the untapped potential of EVs for disease diagnosis and rapid evaluation of treatment efficacy.

Research on EV heterogeneity, particularly studies designed to characterize the biomarker profiles of single EVs, predominantly use nano-flow cytometry and super-resolution microscopy. However, these analyses require sophisticated and expensive equipment and their results can suffer from contamination with other particles or EV subpopulations in the original biospecimen source, reducing their operational simplicity and their sensitivity and accuracy of their results⁴⁶. Here, we propose that a streamlined super-resolution imaging method that uses inexpensive equipment can be employed to rapidly and sensitively analyze EVs directly captured from diagnostic specimens at the single-particle level to accurately evaluate their phenotypic heterogeneity. Our multiplex RCA-ExM assay approach integrates two distinct methods to permit non-disruptive in situ detection of different biomarkers types (protein and miRNA) expressed in different vesicular regions of intact EVs and employs ExM to visualize their respective signals at single-EV resolution. As summarized in Supplementary Table 6, RCA-ExM provides superior multiplexing capacity for simultaneously detecting protein and miRNA biomarkers on EVs while maintaining compatibility with crude biological specimens, addressing critical limitations of current high-throughput methods.

EVs are highly abundant in many biological specimens, with the generally accepted mean EV concentration in the blood of healthy individuals being ~10⁹ EVs/mL⁴⁷. EV concentrations in blood and other biological samples are also frequently elevated in disease and injury conditions, including cancer. However, the percentage of EVs contributed by the diseased or injured tissue is often quite small, complicating the detection of disease-specific EV biomarkers against abundant backgrounds derived from large and heterogenous EV populations present in diagnostic specimens.

Current methods used to analyze single EVs are also prone to interference from nonspecific vesicles and soluble protein and nucleic acid biomarkers. Moreover, ~40-50% of EVs from tumor cell supernatants lack relevant markers, and the low surface area and volume of these vesicles can limit the biomarker expression levels even on biomarker-positive EVs, to further complicate high sensitivity target detection in single-vesicle analyses. In this study, DNA nanocomplexes that bound membrane proteins specifically expressed on a target EV population were used to directly capture and enrich these EVs from biological specimens for specific analysis. SPs recognizing specific cancer-associated EV surface proteins were also employed to provide additional specificity and to allow sensitive detection of the target EV population by serving as substrates for RCA-mediated fluorescent signal production. This approach ensured sensitive and specific labeling of tumor-derived EpCAM⁺ PD-L1⁺ EVs. Incorporating this EV capture and labeling procedure into a chip-based workflow simplified its operation relative to previously reported methods⁴⁸, which require several purification and washing steps to accurately isolate and analyzing specific EVs. Further, we found that a multiplex RCA-ExM assay, which employed three distinct fluorophores to detect signals inducted upon the specific recognition of two EV surface proteins and a miRNA target, could distinguish seven EV subpopulations in on assay sample, and has the potential for further expansion.

Recent advances in ExM have attenuated optical diffraction constraints for image analysis by using sample preparation, achieving super-resolution imaging via physical expansion^49,50. Notably, the liposome-EV fusion and SP/RCA methods we employed to detect specific EV miRNA and surface protein signals were compatible with ExM, since the former approach preserved the membrane integrity of the resulting vesicle while the latter provided multiple hydrogel attachment sites that were employed for subsequent expansion of the RCA-Lipo-EV vesicles. We therefore employed this combined RCA-ExM approach to analyze EVs directly captured from archived plasma specimens of patients with advanced cancer who received either anti-PD-1 or anti-PD-L1 immunotherapy and healthy controls. These cancer patients were primarily diagnosed with advanced lung cancer (51 of 73 cases), since lung cancer – particularly non-small cell lung cancer (NSCLC) – is the principal therapeutic indication for PD-1/PD-L1 checkpoint inhibitors^51,52. Our RCA-ExM analysis of miR-21 expression in EpCAM⁺ PD-L1⁺ EVs found that these results could accurately distinguish individuals in the cancer patient and healthy control groups and accurately stratify cancer patients according to their immunotherapy response.

Post-treatment plasma EV PD-L1 increases detected in a small cohort of individuals with pre- and post-treatment samples available for analysis may arise from a feedback mechanism where PD-L1 expression is enhanced by IFN-γ signaling driven by T-cell activation⁵³. Notably, the single cancer patient with a partial response to immunotherapy and longitudinal samples revealed decreased miR-21 expression in their EpCAM⁺ PD-L1⁺ EVs, whereas all the remaining individuals with longitudinal samples demonstrated miR-21 increases and progressive disease following immunotherapy. Synchronous miR-21 and PD-L1 increases in these non-responders may promote immune evasion through STAT3-regulated signaling pathways⁵⁴, though direct experimental validation of this mechanism is needed. However, larger cancer patient cohorts with longitudinal samples are required to analyze the heterogeneity in treatment regimens. Nevertheless, these results suggest that the RCA-ExM platform can provide a robust framework for multiplex profiling of EV biomarkers across heterogeneous immunotherapy regimens to address a critical need for rapid monitoring of the efficacy of personalized cancer treatments.

The modular design of the RCA-ExM assay platform should also allow it to be rapidly repurposed to detect and analyze other disease-associated EVs populations by exchanging the aptamers or MBs employed to detect specific protein biomarkers on the EV surface or nucleic acid sequences in the EV lumen. For example, replacing the EpCAM and PD-L1-specific aptamers of the current assay with aptamers that recognize neuro-specific (ITGB1) or inflammation-associated (α4β7 integrin) factors could permit its use in diagnostic applications for Alzheimer’s disease and rheumatoid arthritis, respectively^55,56. However, systematic validation studies should be performed to characterize the diagnostic performance of the core RCA-ExM assay platform across a broad spectrum of cancer types, particularly immunotherapy-responsive malignancies, and chronic disease conditions to confirm this expected adaptability. Extensive confirmation of the diagnostic and prognostic utility of this assay platform with liquid biopsy specimens from a diverse array of disease states, from malignant to chronic inflammatory conditions, should significantly advance its potential for use in personalized medicine approaches tailored to specific patient populations identified by distinct EV biomarker profiles.

Taken together, our results indicate that two non-destructive EV biomarker detection methods can be integrated with ExM to permit the precise visualization and sensitive analysis of single-EV biomarker profiles for EV populations directly captured from plasma specimens. This approach can thus provide detailed information about the molecular heterogeneity of target EV populations at the single-EV level, and how specific EV composition differences may be associated with distinct disease phenotypes. This method could thus serve as a robust tool for EV research and means to improve cancer diagnosis and immunotherapy assessments, and thus contribute to future advances in EV biology and personalized medicine.

Methods

Materials

All oligonucleotides (Supplementary Tables 3 and 4) were purchased from Sangon Biotech Co., Ltd. (Shanghai, China), as were the 1× TE buffer, 10× TM buffer, 1× TBE premixed powder, DNA ladder, 1 M MgCl₂ (DEPC), DTT solution (2 M), glycerol, and 20× SSC used in this study. 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE), and 3β [N- (N’, N’-dimethylaminoethane)-carbamoyl] cholesterol (DC-Chol) were purchased from Sigma Chemical Co. (St. Louis, MO, USA) and 1,1’-dioctadecyl-3,3,3’,3’ tetramethylindocarbocyanine perchlorate (DiI) and 3,3’-dioctadecyloxacarbocyanine perchlorate (DiO) were purchased from Beyotime Biotechnology Co., Ltd (Shanghai, China). The duplex-specific nuclease (DSN) employed to amplify miR-21 signal was obtained from Evrogen Joint Stock Company (Moscow, Russia). Nb.BbvCI, quick T4 DNA ligase, phi29 DNA polymerase and deoxynucleotide (dNTP) solution mix were purchased from New England Biolabs Inc. (Ipswich, MA, USA). Phosphate-buffered saline (PBS), Roswell Park Memorial Institute (RPMI) 1640 cell culture medium, Dulbecco’s modified Eagle’s medium (DMEM) cell culture medium, fetal bovine serum (FBS), and penicillin–streptomycin was purchased from Gibco (USA). The 3-aminopropyltriethoxysilane (APTES), m-maleimidobenzoyl-N-hydroxysuccinimide ester (MBS), ammonium persulfate (APS), and N, N, N’, N’-tetramethylethylenediamine (TEMED) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Polydimethylsiloxane (PDMS; Sylgard 184, Dow 110 Corning) and 4-inch silicon wafers were obtained from Chip Technology Co., Ltd. (Suzhou, China). Sodium acrylate, acrylamide, sodium chloride (NaCl), and N, N’-methylenebisacrylamide were purchased from Aladdin Biochemical Technology Co., Ltd. (Shanghai, China).

Cell culture and EV isolation

A375 (ATCC® CRL-1619™), MCF-7 (ATCC® HTB-22™), and U251 (Chinese Academy of Sciences Cell Bank) cells and Jurkat (ATCC® TIB-152™) cells were cultured in DMEM or RPMI 1640 culture media supplemented with 10% FBS and 1% penicillin–streptomycin, respectively, and cultured at 37 °C in a 5% CO₂ incubator. Cells were cultured until they reached 80%, and then switched to serum-free media reduce contamination from FBS-derived EVs and cultured for 48 h to generate conditioned media that was then processed to isolate the EVs secreted by these cells. Briefly, culture supernatants were centrifuged at 2000 × g for 20 min at 4 °C to remove cells and cellular debris, then centrifuged at 16500 × g for 45 min at 4 °C, and filtered through a 0.22-μm filter to remove larger EVs. This filtrate was then subjected to two rounds of centrifugation at 110000 × g and 4 °C for 70 min each, which the supernatant was discarded and the EV pellet was resuspended in PBS and stored at −80 °C for future use.

Detachable chip preparation and characterization

Detachable chips were generated by the following procedure, where all incubations were performed at room temperature unless specified otherwise. First, a premixed 10:1 (w/w) polymer: crosslinker PDMS solution was poured onto a molded silicon wafer and incubated at 80 °C for 30 min, and the cured PDMS substrate was peeled from the molded silicon wafer and perforated using a puncher. Both this membrane and a clean glass slide were then activated by a 1 min oxygen plasma treatment and assembled to generate a detachable chip. The surface of this chip was then modified by a 1 h incubation with a 4% APTES/ methanol solution (v/v), washed with ethanol, and dried for 1 h at 100 °C. The resulting amino-functionalized surface of this chip was incubated with 1 mM MBS for 1 h, washed with PBS, and immediately incubated with a nanocomplex solution for 2 h.

Functionalized chip-based analysis of capture and enzymatic release efficiency

To measure the capture behavior and Nb.BbvCI-mediated release efficiency of the detachable chips, the wells of amino-functionalized chips were incubated with varying concentrations (0–1 μM) of FAM-labeled nanocomplexes at 25 °C for 2 h, washed three time with PBS, and then analyzed with an inverted fluorescence microscope to measure their pre- digestion fluorescent signal (F_pre). These nanocomplex-conjugated chip wells were then incubated at 37 °C for 2 h with Nb.BbvCI (10 U/μL) in 1× Nicking Buffer (NEB), washed three time with PBS, and analyzed by an inverted fluorescence microscope to measure their post-digestion fluorescent signal (F_post). F_pre and F_post measurements were recorded under identical settings, and release efficiency was calculated as: Release efficiency (%) = (1-[F_pre/F_post]) × 100.

Validation of the switch hairpin probe–mediated RCA strategy

SPs were incubated at 95 °C for 5 min in a thermal cycler (Bio-Rad T100, Hercules, CA, USA), then slowly cooled to 25 °C and held at this temperature for 10 min to allow complete annealing of the complementary sequence in the oligonucleotide, and annealing was then analyzed by agarose gel electrophoresis. Monodisperse, trypsin-digested MCF-7 and A375 cell aliquots (5 × 10⁵ cells) suspended in 200 μL of binding buffer (PBS with 5 mM MgCl₂) supplemented with 200 nM PD-L1-S-L were incubated for 1 h at 37 °C on a vortex mixer, and washed twice with PBS. Cells were then resuspended in PBS and incubated with 500 nM RCA template for 30 min at 37 °C, and then supplemented 0.5 μL of quick T4 DNA ligase (total ligation volume of 50 μL) and incubated for 15 min at 25 °C on a vortex mixer. RCA was initiated by adding phi29 DNA polymerase to the cell suspension and incubation at 37 °C for 10 min, then adding reporter (1 μM final concentration), incubating the sample for 50 min more, after which the cells were either washed twice with PBS, or stained with DiI at 37 °C for 15 min, and then PBS washed to remove unbound dye. Cell samples that were not stained with DiI were suspended in PBS and analyzed by flow cytometry, as described below. DiI-stained cells were imaged using a confocal microscope (Olympus, FV3000, Tokyo, Japan) with a 60× objective.

Flow cytometric analysis of EVs

To enable flow cytometric detection of sub-diffraction-limited EVs, 4-μm aldehyde-functionalized latex beads (Thermo Fisher Scientific, Cat # A37304) were conjugated with 20 μL EV samples through a 15-min incubation at 25 °C followed by the and addition of 250 μL PBS and an additional 2-h incubation. Non-specific binding sites were blocked using 25 μL of PBS-based blocking buffer (20% w/v BSA) after a 30-min incubation. Target-specific labeling was performed by incubating EV-coupled beads with 200 nM FAM-conjugated PD-L1 probe and 200 nM Cy5-modified EpCAM probe in 200 μL labeling buffer (PBS, 0.55 mM MgCl₂, 0.5% BSA, pH 7.4) for 40 min at room temperature. The EV-coated beads where then washed with PBS, after which the EV associated FAM (PD-L1) and Cy5 (EpCAM) fluorescent signal intensity of these beads (1 × 10⁴ events/sample) was measured a CytoFLEX S flow cytometry system.

EV membrane protein labeling and analysis

EV aliquots (20 μL) isolated from culture supernatants were supplemented with 80 μL PBS and added to individual wells of nanocomplex-conjugated chips, incubated for 1 h at 37 °C, then washed three times with PBS. EVs captured in these chip wells were then incubated with 200 nM SP (PD-L1-L-S) in 100 μL PBS buffer supplemented 0.55 nM MgCl₂ for 1.5 h at 37 °C, washed twice with PBS, incubated with 10 μM RCA template for 30 min at 37 °C, and then unbound template was removed by careful aspiration. These samples were resuspended in 5 μL 10× TM buffer and mixed with 25 μL of 2 × T4 DNA ligase reaction buffer and 0.5 μL of quick T4 DNA ligase (400U/μL). The total ligation reaction volume was adjusted to 50 μL with nuclease-free water, then incubated at 25 °C for 15 min. Next samples were supplemented with 10 × phi29 DNA polymerase reaction buffer (10 μL), phi29 DNA polymerase (1 μL, 10 U/mL), and dNTP (10 μL, 10 mM) and incubated for 1 h at 37 °C, with 1 μM Reporter 2 added after a 10-minute incubation period. Following this incubation, the EVs were washed twice with PBS, stained for 15 min at 37 °C with the membrane dye DiI, washed twice with PBS, and then resuspended in PBS and imaged with an inverted fluorescence microscope (IX83, Olympus, Tokyo, Japan). Chip-captured EVs were subsequently treated with 10 U/mL Nb.BbvCI in 1 × Nicking Buffer (NEB) for 2 h at 37 °C, and their wells were then washed three time with PBS, and analyzed for differences in their pre- vs. post-digestion fluorescent signal with an inverted fluorescence microscope.

Preparation of liposomes

Cationic liposomes were prepared using a thin-film hydration method where DOPC, DOPE, and DC-Chol were mixed at a 2:1:1 molar ratio in 1 mL of chloroform in a 100-mL round-bottom flask⁵⁷, before the chloroform was evaporated using rotary evaporator to yield a thin lipid film, which was then stored under vacuum overnight to remove residual chloroform. The resulting lipid film was then sonicated for 30 s with 1 mL PBS and the resulting lipid suspension was then extruded 11 times through 0.2-µm porous polycarbonate membranes using an Avanti mini-extruder (Avanti Polar Lipids, Inc., Alabaster, AL, USA) to produce homogeneous suspensions of uniformly sized liposomes.

Reagent-loaded liposomes were fabricated to contain either MB-21 or DSN/MB-21. MB-21-loaded liposomes were fabricated by hydrating dry lipid film samples with 1 mL PBS containing 4 µM MB-21 for 2 h at 37 °C⁵⁸. Liposomes loaded with DSN/MB-21 were fabricated in the same manner using 5 U of DSN and 10× DSN buffer (500 mM Tris-HCl, 50 mM MgCl₂, 10 mM DTT). Both mixtures were subsequently extruded through a 0.2-µm porous polycarbonate membranes 11 times using an Avanti mini-extruder (Avanti Polar Lipids, Inc., Alabaster, AL, USA) to produce a homogeneous suspension of uniformly sized liposomes. A stock DSN solution was generated by dissolving DSN in a 25-mM Tris-HCl solution (pH 8.0)/50% glycerol and stored at −20 °C until use.

Characterization of EVs, liposomes and Lipo-EVs

Lipo-EVs generated for characterization studies were produced by mixing cationic liposomes and A375 EVs at 1:1 ratio and incubating these samples at 37 °C for 2 h. All size distributions and zeta potentials were measured using a Zetasizer nano 2S instrument (Malvern). TEM images of EV, cationic liposome and Lipo-EV samples were captured by pipetting a 20 μL aliquot of these sample onto a bronze mesh, which was allowed to settle for 1 min, after which this mesh was incubated with 5 μL 2% uranyl acetate for 1 min at 25 °C, washed with ultrapure water, air dried for 30 min in a fume hood, and then imaged using a JEM-2010HR microscope.

FRET-based detection of EV-Liposome fusion via DiO fluorescence recovery

EV–liposome fusion was investigated using a FRET assay that detected reduced DiI and increased DiO signal when EVs labeled with both dyes fused with unlabeled liposomes to increase vesicle surface area and the distance and energy transfer between the FRET donor (DiO) and acceptor (DiI) dyes. Briefly, EV aliquots EVs (20 μL) were incubated at 37 °C for 30 min with DiO and DiI (20-μM each) to generate DiO/DiI double-labeled EVs, after which these samples were applied to 100 kDa filters (Amicon) that were centrifuged three times for 20 min at 4000 × g, with two PBS washes, to remove free dye from the retained double-labeled EVs. Double-labeled EVs samples were incubated with 10 μL of liposomes at 37 °C for 2 h and analyzed every 10 min using a microplate reader (BioTek SynergyH1) to quantify an increase in DiO signal (510 nm) consistent increased EV-liposome fusion events and reduced DiO-to-DiI energy transfer.

Validation of the DSN circuit–mediated signal amplification strategy for miR-21 detection

EVs aliquots (20 μL) were incubated with free MB-21, or MB-21-loaded and DSN/MB-21-loaded liposomes for 2 h at 37 °C, respectively. Corresponding EV-free negative control groups were set up in parallel. After incubation, the Cy3 (the fluorophore in MB-21) emission spectra (excitation: 530 nm; emission range: 555–700 nm) were recorded using a FluoroMax-4 fluorescence spectrometer (Horiba Scientific, USA).

Synthesis and imaging of RCA-Lipo-EVs

EV aliquots (20 μL) isolated from cell culture supernatants were supplemented with 80 μL PBS and added to individual wells of nanocomplex-conjugated chips, incubated for 1 h at 37 °C, then washed three times with PBS. EVs captured in these chip wells were then incubated with 200 nM SP (PD-L1-L-S) in 100 μL PBS buffer supplemented 0.55 nM MgCl₂ for 1.5 h at 37 °C, and washed twice with PBS. Sample wells were then incubated with MB/DSN-loaded liposomes for 2 h at 37 °C, then washed with PBS and incubated with 10 μM RCA template for 30 min at 37 °C, and unbound template was removed by careful aspiration. These samples were resuspended in 5 μL 10× TM buffer, mixed with 25 μL of 2× T4 DNA ligase reaction buffer and 0.5 μL of quick T4 DNA ligase (400U/μL) adjusted to 50 μL with nuclease-free water, and incubated at 25 °C for 15 min. Next samples were supplemented with 10× phi29 DNA polymerase reaction buffer (10 μL), phi29 DNA polymerase (1 μL, 10 U/mL), and dNTP (10 μL, 10 mM) and incubated for 1 h at 37 °C, with 1 μM Reporter 2 added after a 10-minute incubation period. Following this incubation, the sample wells were washed twice with PBS and the RCA-Lipo-EVs were imaged with an inverted fluorescence microscope (IX83, Olympus, Tokyo, Japan).

Hydrogel embedding of RCA-Lipo-EVs

RCA-Lipo-EVs captured on the reaction chips (see above) were released by a 2 h incubation with Nb.BbvCI (10 U/μL) at 37 °C in Nicking Enzyme Buffer (NEB), after which wells were supplemented with the hydrogel monomer solution (Supplementary Table 5)⁴⁹ to allow the reporter acrydite groups to react with the hydrogel monomers. Hydrogel formation was then initiated by the addition of 2.5 μL of 10% TEMED (w/w) and 2.5 μL 10% ammonium persulfate [w/w] to achieve a final concentration of 0.2% (w/w) for both. The chip wells were then covered with a cover slide to reduce air exposure, and incubated overnight in a 4 °C refrigerator to allow complete hydrogel polymerization. Hydrogels containing embedded vesicles were then transferred to petri dishes and incubated for 30 min in four successive changes of 0.05× SSC buffer to induce vesicle swelling required for single-vesicle imaging by a conventional inverted fluorescence microscopy system. RCA-Lipo-EV images captured before and after hydrogel expansion were captured using a 40× objective. For pre-expansion hydrogel samples, a 10× objective was used to capture a reference image that as a map to select regions of interest for capture with a 40× objective. For post-expansion hydrogel samples, a 10× objective was used to find regions corresponding to pre-expansion images, which were then imaged with a 40× objective. The time-dependent vesicle swelling ratio Q(t) was measured by evaluating vesicle diameter changes before and after each buffer change in the hydrogel expansion process, using the equation Q(t) = D(t)/D₀, where D₀ and D(t) represent hydrogel diameters detected at time 0 and subsequent time points, respectively.

Cell specificity of RCA-Lipo-EV results

The specificity of our multi-bomarker phenotyping approach was analyzed using an assay that employed a CD63 specific allosteric aptamer to capture target EVs, two EpCAM- and PD-L1-specifc SPs (EpCAM-S-L2 and PD-L1-S-L) and an miR-21-specific MB (MB-21). EVs isolated from A375, U251, MCF-7, and Jurkat cell cultures were individually captured in wells conjugated with CD63-specific aptamer/SH-ssDNA nanocomplexes, EVs captured in aptamer-nanocomplex conjugated wells were processed through target labeling (200 nM PD-L1-S-L in PBS + 0.55 mM MgCl₂; 37 °C, 1.5 h), liposome conjugation (MB/DSN-loaded liposomes; 37 °C, 2 h) and RCA amplification (10 μM template + phi29 DNA polymerase (10 U/mL, 37 °C, 1 h), to determine if the biomarker profiles of these EVs derived from these cell lines matched the reported biomarker profiles of their parental cells. Full protocols with buffer formulations in Methods section “Synthesis and imaging of RCA-Lipo-EVs”.

The ability of this approach to resolve distinct EV populations was analyzed by mixing EVs from these four cell lines and capturing this mixture on a nanocomplex-modified chip that used the same CD63 aptamer-based EV capture strategy. These EVs were then incubated with a mixture of the two SPs (EpCAM-S-L2 and PD-L1-S-L) that was pre-optimized for simultaneous detection of these surface markers, then subjected to RCA, fusion with the MB-21-loaded cationic liposomes, and RCA-ExM imaging to resolve signal profiles from distinct EV subsets.

Clinical sample collection and analysis

Blood samples collected in EDTA were obtained from cancer patients or healthy donors were obtained seen at the First Affiliated Hospital of Sun Yat-sen University who signed informed consent forms that allowed research use of these samples and their anonymized data. These EDTA blood samples were centrifuged at 2000 × g for 10 min to obtain plasma specimens, which were transferred to new 1.5-ml Eppendorf tubes and centrifuged for 20 min at 10,000 × g and 4 °C to remove large vesicles. The resulting supernatants were then transferred to new microcentrifuge tubes and stored at −80 °C until use. RCA-ExM analyses performed on clinical samples used 20 µL samples volumes and the same analysis procedure was used with EVs from cell culture supernatants. All sample collection and data sharing procedures were approved by the Institutional Review Board (IRB) of Sun Yat-sen University (IRB approval number 2019L054-1).

Statistical analysis

Mean fluorescence intensity (MFI) was quantified using NIH Image J software. The Pearson’s correlation coefficient (r) for fluorescence co-localization was determined with Fiji software. All flow cytometry data were analyzed with Flow Jo software (Version 10). P values were determined by using unpaired two-tailed t-tests or one-way ANOVA (two-sided) with Tukey’s multiple comparisons test. Origin (version 9.0), GraphPad Prisim (version 10.2.1) were used for data analyzing and graph presentation. Python (version 3.9) was used for rose chart. R (version 4.3.2) was used for hierarchical clustering visualization and ROC curve.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The main data supporting the results in the study are available within the paper and its Supplementary Information. Source data is available for Figs. 2–5 and Supplementary Figs. 2–4, 6, 7, 9, 10, 12–15, 19 and 20 in the associated source data file. The raw data has been provided in source data and has been deposited in Figshare database (https://doi.org/10.6084/m9.figshare.29205071). Additional relevant information is available from the corresponding author. Source data are provided with this paper.

References

Atay, S. et al. Oncogenic KIT-containing exosomes increase gastrointestinal stromal tumor cell invasion. Proc. Natl. Acad. Sci. USA 111, 711–716 (2014).
Article ADS CAS PubMed Google Scholar
Wang, H. et al. A Self-Serviced-Track 3D DNA Walker for Ultrasensitive Detection of Tumor Exosomes by Glycoprotein Profiling. Angew. Chem. Int Ed. Engl. 61, e202116932 (2022).
Article CAS PubMed Google Scholar
Kalluri, R. & LeBleu, V. S. The biology, function, and biomedical applications of exosomes. Science 367, eaau6977 (2020).
Article CAS PubMed PubMed Central Google Scholar
Deng, J. et al. One-step thermophoretic AND gate operation on extracellular vesicles improves diagnosis of prostate cancer. Angew. Chem. Int Ed. Engl. 61, e202207037 (2022).
Article CAS PubMed Google Scholar
Lin, C. et al. Localized plasmonic sensor for direct identifying lung and colon cancer from the blood. Biosens. Bioelectron. 211, 114372 (2022).
Article CAS PubMed Google Scholar
Wen, S. W. et al. A. Moller, breast cancer-derived exosomes reflect the cell-of-origin phenotype. Proteomics 19, e1800180 (2019).
Article PubMed Google Scholar
Zhou, H. et al. Liquid biopsy at the frontier of detection, prognosis and progression monitoring in colorectal cancer. Mol. Cancer 21, 86 (2022).
Article CAS PubMed PubMed Central Google Scholar
Poggio, M. et al. Suppression of exosomal PD-L1 induces systemic anti-tumor immunity and memory. Cell 177, 414–427 e413 (2019).
Article CAS PubMed PubMed Central Google Scholar
Daassi, D., Mahoney, K. M. & Freeman, G. J. The importance of exosomal PDL1 in tumour immune evasion. Nat. Rev. Immunol. 20, 209–215 (2020).
Article CAS PubMed Google Scholar
Li, B., Cao, Y., Sun, M. & Feng, H. Expression, regulation, and function of exosome-derived miRNAs in cancer progression and therapy. FASEB J. 35, e21916 (2021).
Article CAS PubMed Google Scholar
Li, J., He, D., Bi, Y. & Liu, S. The emerging roles of exosomal miRNAs in breast cancer progression and potential clinical applications. Breast Cancer (Dove Med Press) 15, 825–840 (2023).
CAS PubMed Google Scholar
Liu, H., Chen, C., Zeng, J., Zhao, Z. & Hu, Q. MicroRNA-210-3p is transcriptionally upregulated by hypoxia induction and thus promoting EMT and chemoresistance in glioma cells. PLoS One 16, e0253522 (2021).
Article CAS PubMed PubMed Central Google Scholar
Liu, L. et al. Hepatocyte-derived extracellular vesicles miR-122-5p promotes hepatic ischemia reperfusion injury by regulating Kupffer cell polarization. Int Immunopharmacol. 119, 110060 (2023).
Article CAS PubMed Google Scholar
Zheng, S., Liao, J., Sun, M., Liu, R. & Lv, J. Extracellular shuttling miR-21 contributes to esophageal cancers and human umbilical vein endothelial cell communication in the tumor microenvironment and promotes tumor angiogenesis by targeting phosphatase and tensinhomolog. Thorac. Cancer 14, 3119–3132 (2023).
Article CAS PubMed PubMed Central Google Scholar
Gardiner, C. et al. Techniques used for the isolation and characterization of extracellular vesicles: results of a worldwide survey. J. Extracell. Vesicles 5, 32945 (2016).
Article PubMed Google Scholar
Kowal, J., Tkach, M. & Thery, C. Biogenesis and secretion of exosomes. Curr. Opin. Cell Biol. 29, 116–125 (2014).
Article CAS PubMed Google Scholar
Kowal, J. et al. Proteomic comparison defines novel markers to characterize heterogeneous populations of extracellular vesicle subtypes. Proc. Natl. Acad. Sci. USA 113, E968–E977 (2016).
Article CAS PubMed PubMed Central Google Scholar
Tkach, M. & Thery, C. Communication by extracellular vesicles: where we are and where we need to go. Cell 164, 1226–1232 (2016).
Article CAS PubMed Google Scholar
Bordanaba-Florit, G., Royo, F., Kruglik, S. G. & Falcon-Perez, J. M. Using single-vesicle technologies to unravel the heterogeneity of extracellular vesicles. Nat. Protoc. 16, 3163–3185 (2021).
Article CAS PubMed Google Scholar
de Voogt, W. S., Tanenbaum, M. E. & Vader, P. Illuminating RNA trafficking and functional delivery by extracellular vesicles. Adv. Drug Deliv. Rev. 174, 250–264 (2021).
Article PubMed Google Scholar
Ning, B. et al. Liposome-mediated detection of SARS-CoV-2 RNA-positive extracellular vesicles in plasma. Nat. Nanotechnol. 16, 1039–1044 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Lei, Y. et al. Simultaneous subset tracing and miRNA profiling of tumor-derived exosomes via dual-surface-protein orthogonal barcoding. Sci. Adv. 9, eadi1556 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Li, Y. et al. Molecular identification of tumor-derived extracellular vesicles using thermophoresis-mediated DNA computation. J. Am. Chem. Soc. 143, 1290–1295 (2021).
Article CAS PubMed Google Scholar
Wu, J. et al. Identifying the phenotypes of tumor-derived extracellular vesicles using size-coded affinity microbeads. J. Am. Chem. Soc. 144, 23483–23491 (2022).
Article CAS PubMed Google Scholar
Thakur, B. K. et al. Double-stranded DNA in exosomes: a novel biomarker in cancer detection. Cell Res 24, 766–769 (2014).
Article CAS PubMed PubMed Central Google Scholar
Lotvall, J. et al. Minimal experimental requirements for definition of extracellular vesicles and their functions: a position statement from the international society for extracellular vesicles. J. Extracell. Vesicles 3, 26913 (2014).
Article PubMed Google Scholar
van der Pol, E. et al. Particle size distribution of exosomes and microvesicles determined by transmission electron microscopy, flow cytometry, nanoparticle tracking analysis, and resistive pulse sensing. J. Thromb. Haemost. 12, 1182–1192 (2014).
Article PubMed Google Scholar
Kyriazi, M. E. et al. Multiplexed mRNA sensing and combinatorial-targeted drug delivery using DNA-gold nanoparticle dimers. ACS Nano 12, 3333–3340 (2018).
Article CAS PubMed Google Scholar
Mao, C. et al. Feature-rich covalent stains for super-resolution and cleared tissue fluorescence microscopy. Sci. Adv. 6, eaba4542 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Wang, W. et al. Live-cell imaging and analysis reveal cell phenotypic transition dynamics inherently missing in snapshot data. Sci. Adv. 6, eaba9319 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Wu, J. et al. Live imaging of mRNA using RNA-stabilized fluorogenic proteins. Nat. Methods 16, 862–865 (2019).
Article CAS PubMed PubMed Central Google Scholar
Huang, B., Babcock, H. & Zhuang, X. Breaking the diffraction barrier: super-resolution imaging of cells. Cell 143, 1047–1058 (2010).
Article CAS PubMed PubMed Central Google Scholar
Schnitzbauer, J., Strauss, M. T., Schlichthaerle, T., Schueder, F. & Jungmann, R. Super-resolution microscopy with DNA-PAINT. Nat. Protoc. 12, 1198–1228 (2017).
Article CAS PubMed Google Scholar
Shim, S. H. et al. Super-resolution fluorescence imaging of organelles in live cells with photoswitchable membrane probes. Proc. Natl. Acad. Sci. USA 109, 13978–13983 (2012).
Article ADS CAS PubMed PubMed Central Google Scholar
Yan, R., Wang, B. & Xu, K. Functional super-resolution microscopy of the cell. Curr. Opin. Chem. Biol. 51, 92–97 (2019).
Article CAS PubMed Google Scholar
Chen, F., Tillberg, P. W. & Boyden, E. S. Optical imaging. Expansion microscopy. Science 347, 543–548 (2015).
Article ADS CAS PubMed PubMed Central Google Scholar
Lin, B. et al. Tracing tumor-derived exosomal PD-L1 by dual-aptamer activated proximity-induced droplet digital PCR. Angew. Chem. Int Ed. Engl. 60, 7582–7586 (2021).
Article CAS PubMed Google Scholar
Martowicz, A., Spizzo, G., Gastl, G. & Untergasser, G. Phenotype-dependent effects of EpCAM expression on growth and invasion of human breast cancer cell lines. BMC Cancer 12, 501 (2012).
Article CAS PubMed PubMed Central Google Scholar
Kapranov, P. et al. RNA maps reveal new RNA classes and a possible function for pervasive transcription. Science 316, 1484–1488 (2007).
Article ADS CAS PubMed Google Scholar
Wang, Z., Gerstein, M. & Snyder, M. RNA-Seq: a revolutionary tool for transcriptomics. Nat. Rev. Genet 10, 57–63 (2009).
Article CAS PubMed PubMed Central Google Scholar
Fan, T. et al. Label-free fluorescence detection of circulating microRNAs based on duplex-specific nuclease-assisted target recycling coupled with rolling circle amplification. Talanta 200, 480–486 (2019).
Article CAS PubMed Google Scholar
Chang, X. et al. Construction of a multiple-aptamer-based DNA logic device on live cell membranes via associative toehold activation for accurate cancer cell identification. J. Am. Chem. Soc. 141, 12738–12743 (2019).
Article CAS PubMed Google Scholar
Went, P. T. H. et al. Frequent EpCam protein expression in human carcinomas. Hum. Pathol. 35, 122–128 (2004).
Article CAS PubMed Google Scholar
Ren, Y. et al. Multivalent DNA flowers for high-performance isolation, detection, and release of tumor-derived extracellular vesicles. ACS Appl Mater. Interfaces 15, 55358–55368 (2023).
Article CAS PubMed Google Scholar
Bano, R., Ahmad, F. & Mohsin, M. A perspective on the isolation and characterization of extracellular vesicles from different biofluids. RSC Adv. 11, 19598–19615 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Dong, L. et al. Comprehensive evaluation of methods for small extracellular vesicles separation from human plasma, urine and cell culture medium. J. Extracell. Vesicles 10, e12044 (2020).
Article CAS PubMed Google Scholar
Ferguson, S. & Weissleder, R. Modeling EV kinetics for use in early cancer detection. Adv. Biosyst. 4, e1900305 (2020).
Article PubMed PubMed Central Google Scholar
Oeyen, E. et al. Ultrafiltration and size exclusion chromatography combined with asymmetrical-flow field-flow fractionation for the isolation and characterisation of extracellular vesicles from urine. J. Extracell. Vesicles 7, 1490143 (2018).
Article PubMed PubMed Central Google Scholar
Li, H., Warden, A. R., He, J., Shen, G. & Ding, X. Expansion microscopy with ninefold swelling (NIFS) hydrogel permits cellular ultrastructure imaging on conventional microscope. Sci. Adv. 8, eabm4006 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Klimas, A. et al. Magnify is a universal molecular anchoring strategy for expansion microscopy. Nat. Biotechnol. 41, 858–869 (2023).
Article CAS PubMed PubMed Central Google Scholar
Reck, M. et al. Pembrolizumab versus chemotherapy for PD-L1–positive non–small-cell lung cancer. N. Engl. J. Med 375, 1823–1833 (2016).
Article CAS PubMed Google Scholar
Topalian, S. L., Taube, J. M., Anders, R. A. & Pardoll, D. M. Mechanism-driven biomarkers to guide immune checkpoint blockade in cancer therapy. Nat. Rev. Cancer 16, 275–287 (2016).
Article CAS PubMed PubMed Central Google Scholar
Garcia-Diaz, A. et al. Interferon receptor signaling pathways regulating PD-L1 and PD-L2 expression. Cell Rep. 19, 1189–1201 (2017).
Article CAS PubMed PubMed Central Google Scholar
Yadav, R. et al. The miRNA and PD-1/PD-L1 signaling axis: an arsenal of immunotherapeutic targets against lung cancer. Cell Death Discov. 10, 414 (2024).
Article CAS PubMed PubMed Central Google Scholar
You, Y. et al. Human neural cell type-specific extracellular vesicle proteome defines disease-related molecules associated with activated astrocytes in Alzheimer’s disease brain. J. Extracell. Vesicles 11, e12183 (2022).
Article CAS PubMed PubMed Central Google Scholar
Xu, P. et al. Intestinal sulfation is essential to protect against colitis and colonic carcinogenesis. Gastroenterology 161, 271–286.e211 (2021).
Article CAS PubMed Google Scholar
Loffler, P. M. G. et al. A DNA-programmed liposome fusion cascade. Angew. Chem. Int Ed. Engl. 56, 13228–13231 (2017).
Article PubMed Google Scholar
Cui, L. et al. A general strategy for detection of tumor-derived extracellular vesicle microRNAs using aptamer-mediated vesicle fusion. Nano Today 46, 101599 (2022).

Download references

Acknowledgements

We would like to acknowledge support by the National Natural Science Foundation of China (82072087 to Y. Z., 31970893 to M. W.), funding by Guangdong Provincial Natural Science Foundation (2024A1515011932 to Y. Z., 2025A1515010007 to Y. Z.), and research funding from the National Institutes of Health (U01CA252965 to T. H.).

Author information

Authors and Affiliations

State Key Laboratory of Anti-Infective Drug Discovery and Development; School of Pharmaceutical Sciences, Sun Yat-sen University, Guangzhou, China
Jiacheng Wu, Quanhao Dou, Miao Mao, Xin Wan & Yuanqing Zhang
Zhongshan School of Medicine, Sun Yat-Sen University, Guangzhou, China
Minhao Wu
Center of Cellular and Molecular Diagnosis, Tulane University School of Medicine, New Orleans, LA, USA
Tony Y. Hu

Authors

Jiacheng Wu
View author publications
Search author on:PubMed Google Scholar
Quanhao Dou
View author publications
Search author on:PubMed Google Scholar
Miao Mao
View author publications
Search author on:PubMed Google Scholar
Xin Wan
View author publications
Search author on:PubMed Google Scholar
Minhao Wu
View author publications
Search author on:PubMed Google Scholar
Tony Y. Hu
View author publications
Search author on:PubMed Google Scholar
Yuanqing Zhang
View author publications
Search author on:PubMed Google Scholar

Contributions

J.W. and Y.Z. conceived the project. J.W. performed the experiments. Q.D. and X.W. contributed to the collection and processing of clinical samples. J.W. and M.M. analyzed the experimental data and interpreted the results. M.W., T.H., and Y.Z. supervised the project. J. W., T.H., and Y.Z. wrote the manuscript. All authors joined in the critical discussion and edited the paper.

Corresponding authors

Correspondence to Minhao Wu, Tony Y. Hu or Yuanqing Zhang.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Reporting Summary

Transparent Peer Review file

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Wu, J., Dou, Q., Mao, M. et al. Single extracellular vesicle imaging via rolling circle amplification–expansion microscopy. Nat Commun 16, 7498 (2025). https://doi.org/10.1038/s41467-025-62613-0

Download citation

Received: 12 January 2025
Accepted: 25 July 2025
Published: 13 August 2025
DOI: https://doi.org/10.1038/s41467-025-62613-0