1. Introduction
Prostate cancer (PC) is one of the most common cancers among men. PC mortality rate is strongly correlated to the stage of the disease when first diagnosed. According to the St. Petersburg population-based cancer registry, the five-year relative survival rate of PC patients drops dramatically with increasing disease stage: 100% (stage I), 100% (stage II), 69% (stage III), and only 24% (stage IV) [
1]. Data arising from other sources including the United States Surveillance, Epidemiology, and End Results (SEER) database demonstrate a similar tendency. This indicates an urgent need to develop a new, effective approach to detect PC earlier and increase the proportion of patients with a good prognosis. Current diagnostics of PC include serum prostate-specific antigen (PSA) determination, digital rectal examination (DRE), transrectal ultrasound (US), and US-guided tissue biopsy [
2]. PSA and related markers, such as (−2)proPSA, and the prostate health index (PHI) have become widely used; however, their wide application did not considerably improve PC diagnostics, mostly due to low specificity. The effectiveness of DRE and US depends on the experience of the physician [
3] and still lack appropriate standardization [
4]. Tissue biopsy is an expensive and invasive procedure. Thus, the discovery of easily testable and highly specific markers of PC would fill the gap between PSA testing and tissue biopsy and would assist with the management of patients with increased PSA levels to improve the overall effectiveness of PC diagnostics.
Circulating with plasma, small extracellular nanovesicles (SEVs), especially exosomes, are recognized as a promising marker of different cancers [
5], including PC [
6]. The involvement of SEVs in PC development and dissemination was comprehensively summarized in recent reviews [
7,
8]. Different components of plasma SEVs, including exosomal RNA and proteins, have been studied as PC markers in dozens of studies (listed in [
8]). Vesicular microRNAs (miRNAs) are the markers of PC most extensively studied over the past few years [
9]. However, the results of these investigations are still not sufficiently reproducible, most likely due to methodological particularities and the fundamental problem of the high heterogeneity of plasma SEVs. The population of plasma SEVs derives from different cells in the bodies in an as-yet unknown ratio. As a small tumor of prostate gland does not alter considerably the composition of the plasma SEV population, the quantification of prostate-derived vesicles with diagnostic purpose would be of little value. The isolation followed by qualitative analysis of the population of prostate-derived SEVs may present a more promising approach for vesicle-based liquid biopsy.
The mature prostate gland is composed of columnar epithelial cells lining the prostate lumen and elongated basal cells separating the lumen from the stroma. PC can arise from both types of cells. Maintaining a certain level of tissue differentiation, transformed epithelial cells of the prostate show a tissue-specific profile of membrane proteins and secrete SEVs bearing tissue-specific markers. For instance, zinc metalloenzyme glutamate carboxypeptidase II (GCPII) or prostate-specific membrane antigen (PSMA) is enriched in prostate epithelial cells and are detected in the membrane of the prostate and prostate-cancer-secreted SEVs [
10,
11]. We hypothesized that a population of PSMA(+)SEVs would much better reflect cancer-specific alterations in the miRNAs occurring in the cells of the prostate gland than the total population of plasma SEVs. To explore this hypothesis, we developed a method to isolate PSMA(+)SEVs using super-paramagnetic beads functionalized by PSMA-specific DNA aptamer [
12] via the click reaction. In combination with the previously described two-phase polymer system for plasma SEV isolation [
13], the proposed technique provides a scaffold for an effective and economic approach to vesicle-based liquid biopsy. We analyzed four PC-associated miRNAs in PSMA(+)SEVs and the total population of plasma vesicles, and compared the diagnostic potency of these tests. The obtained results confirmed that the isolation of PSMA(+)SEVs increases the sensitivity of PC-associated miRNAs analysis and improves the values of the most important diagnostic parameters.
2. Materials and Methods
2.1. Patients
Plasma samples were obtained from patients undergoing treatment at N.N. Petrov National Medical Research Centre of Oncology and Pavlov First Saint Petersburg State Medical University. The study included patients with a histologically confirmed diagnosis of prostate cancer (n = 55) who met the specified criteria: age 56–70 years (median age was 63 years); no chronic or metabolic diseases; prostate cancer stage, T1c–T2c N0 M0; and Gleason score, 5–7. Plasma from healthy male donors (n = 30) age 55–70 years (median age was 61 years) was obtained from the blood transfusion department of the N.N. Petrov National Medical Research Centre of Oncology. Donors and patients gave informed consent to participate in the study.
2.2. Plasma Sampling and Preparation
Blood samples (5 mL) were placed in EDTA tubes. Plasma was immediately separated by centrifugation for 15 min at 1500× g and then stored at −80 °C. Before analysis, the frozen plasma was slowly thawed at 4 °C, sequentially centrifuged at 300, 1500, and 2500× g, and filtered through a 0.2 µm syringe filter (Minisart High Flow, Sartorius, Goettingen, Germany) to remove cellular debris. Plasma prepared by this method is hereafter referred to as pellet pure plasma (PPP) and was used for all experiments.
2.3. Labeling and Isolation of the Total Population of Plasma Extracellular Nanovesicles (SEVs)
The PPP was stained with lipophilic dye and fractionated by size-exclusion chromatography (SEC) as recently described [
14]. Briefly, 2 mL of PPP was mixed with 2 μL of Vybrant™ Dil Cell-Labeling Solution dye solution (50 μM in DMSO; Thermo Fisher Scientific, Walthman, MA, USA) incubated at 37 °C for 20 min under moderate stirring. Each sample of stained plasma was loaded into the SEC column (HansaBioMed, Tallin, Estonia), and 23 fractions of 500 µL were collected in according to the producer’s manual. Fractions 9–11 were mixed and used for further experiments.
2.4. Total EVs Isolation by Two-Phase Polymer System
The total EV population was isolated from PPP using a two-phase polymer system as previously described [
13], with slight modifications. Briefly, a solution of dextran (450–650 kDa, 1.5%) and polyethylene glycol (35 kDa, 3.5%) (both from Sigma-Aldrich, St. Louis, MO, USA) was prepared in the required volume of PPP. To prepare a protein-depleting solution (PDS), the same amounts of polymers were dissolved in an equal volume of phosphate-buffered saline (PBS). The PPP-based solution was centrifuged for 10 min at 1000×
g to achieve phase separation. The upper phase, containing plasma proteins, was replaced with PDS, then the solution was well-mixed and centrifuged again. The upper phase formed during the second round of centrifugation was removed. The lower phase containing SEVs was resuspended in 100 μL of PBS.
2.5. Nanoparticle Tracking Analysis (NTA)
The size and concentration of the isolated vesicles were measured using a NanoSight NS300 analyzer (Malvern Panalytical, Malvern, UK) at camera level: 10, shutter slider: 696, slider gain: 55, and threshold level: 5. Each sample was pumped through the observation camera to record 5 measurements for 30 s for 749 frames at different microvolumes of the same sample. Based on the results of five measurements, the average values of the size and concentration of the nanoparticles in the suspension were calculated. The data were processed in Nanosight NTA 3.4.
2.6. Transmission Cryo-Electron Microscopy (Cryo-TEM)
Visualization was performed using cryo-electron microscopy on a JEOL JEM-1400 transmission electron microscope (JEOL Ltd., Tokyo, Japan) at the Research Resource Center for Molecular and Cell Technologies of Saint-Petersburg State University. The SEV samples at a concentration of 1012 particles/mL were deposited on carbon-coated copper mesh/lacey carbon-supported copper grids, 50 nm in size (Sigma-Aldrich, St. Louis, MO, USA). The excess sample was removed with filter paper. After, the sample was immersed in liquid ethane for rapid freezing and transferred to a cryostat for subsequent analysis using a cryo-microscope.
2.7. Bead-Assisted Flow Cytometry of the Total SEV Population
Plasma SEV surface markers were detected by flow cytometry using an Exo-FACS kit (HansaBioMed, Tallinn, Estonia) according to the manufacturer’s protocol. SEVs were absorbed nonspecifically to the surface of latex microparticles. Tetraspanins CD9 and CD63 were labelled using antibodies conjugated with fluorescent labels PE (CD9-PE, 312105, BioLegends, San Diego, CA, USA) and FITC (CD63-FITC, Ab18235, Abcam, Cambridge, UK). Measurements were performed on a Cytoflex analyzer; data were processed using CytExpert software (both from Beckman Coulter, Brea, CA, USA).
2.8. Dot Blotting
SEVs were isolated from the PPP (2 mL) of healthy donors with a two-phase polymer system and dissolved in 100 µL of PBS. SEV samples (0.4 µL) were applied to a nitrocellulose membrane and allowed to dry. The membrane was incubated in blocking buffer containing 5% bovine serum albumin (BSA) in tris-buffered saline (TBS) at room temperature for 20 min under moderate stirring. Membranes with spotted samples were incubated with primary antibodies against PSMA (133579, Abcam, Cambridge, UK), CD9 (312102, BioLegend, San Diego, CA, USA), CD63 (353039, BioLegend, San Diego, CA USA), and HSP70 (kindly provided by the author of Patent RF, 2722398) at an equivalent concentration of 1 ng/mL in 0.2% BSA solution at 4 °C overnight. Unbound primary antibodies were washed with TBS. Membranes were incubated first in a blocking buffer (inactivated plasma centrifuged for 5 min at 17,000× g and filtered through a 0.1 µm syringe filter), then with secondary antibodies, either goat anti-mouse IgG (6789, Abcam, Cambridge, UK) or goat anti-rabbit IgG (7171, Abcam Cambridge, UK) at a concentration of 0.13 ng/mL in 0.1% BSA solution at 4 °C for 20 min. Finally, membranes were washed twice in TBS and evaluated with PierceTM ECL Western Blotting Substrate and the iBrightTM FL1500 Imaging System (both from ThermoFisher Scientific, Walthman, MA, USA).
2.9. Formation of Complexes for PSMA(+)SEVs Sorption
The complex was composed of super-paramagnetic beads (SPMBs) with the surface functionalized by –N
3 groups (Click Chemistry Tools, Phoenix, AZ, USA) and a PSMA-specific DNA-aptamer (PSMA–Apt) modified by a dibenzocyclooctyne group (DBCO) at the −5′ end. The sequence of the PSMA-specific aptamer (DBCO-5′-GAA TTC GCG TTT TCG CTT TTG CGT TTT GGG TCA TCT GCT TAC GAT AGC AAT GCT-3′) was adopted from [
12] and synthesized by Syntol Ltd., Moscow, Russia. Complex SPMB–PSMA–Apt was formed by the reaction of azide-alkyne cycloaddition. First, 1 µL of SPMB (1 mg/mL) was incubated in a 200 µL I-Block™ Protein-Based Blocking Reagent (T2015, ThermoFisher Scientific, Walthman, MA, USA) at 4 °C for 1 h to prevent nonspecific binding. The SPMBs were then washed twice in 200 µL of PBS and resuspended in 100 µL of PBS. Then 1 µL of PSMA–Apt solution (100 pM) was added to the suspension of blocked SPMB, mixed, and incubated for 3 h at room temperature. The SPMB–Apt complexes were washed twice with 200 µM of PBS to remove the unbound aptamers. To test the efficacy of the formed complexes, we used Cy5.5-labelled aptamer. In this case, the pellet of SPMB–Apt was resuspended in 100 µM of PBS and used for flow cytometry. To capture PSMA(+)SEVs, the pellet of SPMB functionalized by unlabeled aptamer was mixed with the suspension of SEVs.
2.10. PSMA(+)SEVs Sorption by the SPMB–Apt Complex
As described above, 100 µL of SEVs suspension was added to the pellet of SPMB-Apt complexes, mixed, incubated at 4 °C overnight under moderate stirring, washed twice in 200 µL of unbound SEVs, and resuspended in 100 μL of PBS. To assay the efficacy of PSMA(+)SEVs captured by flow cytometry, plasma SEVs were labeled and isolated by SEC as described in
Section 2.3. To assay the efficacy of PSMA(+)SEVs captured by dot blotting, plasma SEVs were isolated with a two-phase polymer system, as described in
Section 2.4.
2.11. Formation of Immuno-Beads for PSMA(+)SEVs Immune-Sorption
To isolate PSMA(+)SEVs, we used the complex of super-paramagnetic beads with the surface functionalized by streptavidin (SPMB-St, K0180, Sileks, Moscow, Russian Federation) and PSMA-specific antibodies (Ab-PSMA, PSMCC8, HyTest Ltd., Turku, Finland) modified by biotin. The complex of SPMB-St–PSMA-Ab was formed as previously described [
15]. Briefly, the SPMB-St was washed 3 times in 200 µL of PBS. Then, 0.5 µL of PSMA-Ab (3.4 mg/mL) was mixed with 1 µL of SPMB-St (1 mg/mL) in 100 µL of PBS and incubated at 4 °C for 1 h. Then, unbound Ab was washed twice in 200 µL of PBS; SPMB-St–PSMA-Ab complexes were used in the following experiments.
2.12. MiRNA Isolation
RNA from the SEVs was isolated with an RNAGEM kit (MicroGem, Dunedin, New Zealand) according to the manufacturer’s protocol. Both the total population of SEVs and PSMA(+)SEVs were isolated from 2 mL of PPP. The total population of SEVs was isolated by a two-phase polymer system. PSMA(+)SEVs were isolated by SPMB–PSMA–Apt complexes. Proteolysis was performed in 50 µL of water solution containing 5 µL of 10× BLUE buffer and 0.2 µL of RNAGEM reagent at 75 °C for 5 min. The concentration of the RNA was estimated using a Qubit 1.0 fluorimeter.
2.13. RT-PCR
The isolated RNA was analyzed by reverse transcription (RT), followed by real-time polymerase chain reaction (PCR) using a commercial ALMIR kit: AL145-5p, AL221-3p, AL451a-3p, AL141-3p (Algimed Techno, Minsk, Belarus) according to the manufacturer’s protocol. RT-reaction mix contained 2 μL of total RNA solution (
Section 2.11), RT primer (50 pM), M-MLV reverse transcriptase (40 U) and proprietary RT buffer in a final volume 10 μL. The RT reaction was conducted for 45 min at 25 °C, then for 5 min at 85 °C to inactivate the enzymatic activity of reverse transcriptase. The following PCR was performed in a reaction volume of 20 μL including: RT reaction mix (4 μL), forward and reverse primers (each for 250 pM), FAM-labelled probe (150 pM) and 10 μL of 2xPCR master mix. Conditions of PCR: 5 min—95 °C, then 40 cycles: 5 s—95 °C; 15 s—60 °C. Any reaction was conducted in technical duplicate.
2.14. Statistics
The experimental data were processed using ImageJ, SigmaPlot 12.0, GraphPad Prizm 8, OriginPro 9.1, and CFX Manager Software 3.1. Statistical differences between groups of samples were evaluated using the nonparametric Mann–Whitney test. ROC analysis was used to assess the diagnostic significance of the developed method.
4. Discussion
Neoplastic transformation induces specific alterations in the cellular miRNA profile. These alterations are supposed to be translated into the composition of miRNAs released from cells into the extracellular space within SEVs. However, the total SEV population is extremely heterogeneous; so, small amounts of tissue-specific vesicles do not significantly affect the overall profile of circulating vesicles. Thus, for diagnostics, it is more considerable to use specific fractions of microvesicles. To create a diagnostic method, SEVs secreted by cells of certain tissue must be separated to detect cancer-indicative alterations in the miRNA content of these vesicles. First, we performed a proof-of-principle study by analyzing the miRNAs in SEVs bearing thyroid peroxidase and revealed the potent ability of this test to discriminate follicular adenoma and follicular cancer of the thyroid gland [
15]. For prostate cancer, the effectiveness of the analysis of specific fractions of SEVs has already been highlighted in the work of Chiara Foroni et al. [
30]. However, in this study, the vesicles with undisclosed cancer-specific surface marker were isolated by immune-capturing and multiple PC-relevant alterations of the androgen receptor were profiled in SEV-associated DNA. In our study, we explored other tactics and a different strategy. We aimed to isolated prostate-derived SEVs instead of cancer-specific vesicles; and we explored aptamer-based technology instead of antibody-based immune-capturing. We supposed that PC-associated alterations of miRNAs in prostate-derived SEV will have diagnostic potency. The obtained results confirmed our hypothesis; however, certain issues requiring further research were highlighted.
The first unexpected result was the high relative amount of PSMA(+)SEVs in plasma. The dot blot analysis cannot distinguish the signal from a few vesicles enriched by the marker of interest from a high number of vesicles with a low amount of the marker. Comparative analysis of certain markers’ expression in vesicles separated by stochastic optical reconstruction microscopy (STORM) would help resolve this issue and will be applied in the future. We suppose that PSMA is expressed at a similar or lower level than CD63 in prostate-derived SEVs, and the obtained ratio of PSMA to CD63 signal intensity reflects the ratio of PSMA(+) to CD63(+) vesicles. This interpretation of the results allows us to state that PSMA(+)SEVs constitute 20–24% of the total vesicular population. If this is true, PSMA(+)SEVs cannot be secreted exclusively by prostate cells. Thus, the fraction of PSMA(+)SEVs is not prostate-specific but rather enriched by prostate-derived SEVs. This conclusion justifies further optimization of the isolating technology using other prostate-specific markers and their combination.
The second interesting result of our investigation is the higher efficacy of complexes functionalized by DNA aptamers compared to complexes functionalized by antibodies. To explain this difference, we assumed similar affinity of both ligands to PSMA and calculated the number of ligand molecules apparently attached to the same amount of SPMBs. Thus, SPMBs were functionalized with PSMA–Apt in 100 μL of reaction mix containing 1.7 × 10−6 g of DNA aptamers with a molecular mass of 17 kDa, which corresponds to 60.2 × 1012 molecules per reaction. Likewise, SPMBs were functionalized with PSMA-Ab in a 100 μL of reaction mix containing 1.7 × 10−6 g of antibodies with a molecular mass of 150 kDa, which corresponds to 6.8 × 1012 molecules per reaction. This calculation indicates an eight-times higher density of SPMA binding sites on the surface of SPMB–PSMA–Apt complexes than on the surface of SPMB–PSMA-Ab complexes. Excluding the possible differences in the efficiency of the binding of ligands to SPMB and the affinity of interaction of ligands with PSMA, this calculation can, in part, explain the higher PSMA(+)SEVs binding capacity of SPMB–PSMA–Apt complexes and may stimulate further development of aptamer-based technology. Also, to prove the effectiveness of using the SPMB-PSMA-Apt complex (via flow cytometry), for determining significant results, magnetic particles without PSMA-aptamer incubated with vesicles labeled with Vibrant-Dil were used as a limit of detection and negative control. It would be more correct to use the lowest concentration of PSMA(+) vesicles that can be detected by a developed method as a limit of detection. Unfortunately, there is no method available to determining this concentration. Therefore, we are planning to continue working on the development of a method to estimate the concentration of PSMA (+) SEVs.
Finally, we want to indicate the most important issue that requires extensive further investigation to advance the proposed technology from bench to bedside. In our study, we empirically selected four PC-related miRNAs and revealed a method to improve their diagnostic potency. We considered that the cancer-related expression alterations of these molecules translate to the miRNA content of prostate-derived SEVs. However, it was reported that the packaging of miRNAs into exosomes or other vesicles is an active and regulated process, and the cellular miRNA pattern is not the same as the vesicular miRNA pattern. This indicates the importance of deeper analysis and miRNA profiling of prostate-derived SEVs isolated from the plasma of healthy donors and PC patients. The selection of highly relevant miRNA markers would increase the overall diagnostic potency of the proposed method.