Monitoring Glucocorticoid Receptor in Plasma-derived Extracellular Vesicles as a Marker of Resistance to Androgen Receptor Signaling Inhibition in Prostate Cancer

Cell Line Cultures and Reagents

The LNCaP human prostate cancer cell line was maintained in RPMI medium supplemented with 10% FBS and 1% penicillin/streptomycin. LREX cells were kindly provided by Charles Sawyers Lab and were maintained in RPMI with l-glutamine and HEPES buffer supplemented with 20% FBS (Omega Scientific Inc) and 1% penicillin/streptomycin. LREX cells were cultured continually in 1 µmol/L ENZA. All cell lines were maintained at 37°C and 5% CO₂ and grown to 70% confluency. Routine testing for Mycoplasma contamination was performed. Cell line purity was confirmed by fingerprinting analysis by the MD Anderson Cancer Center Core Facility in Department of Systems Biology. enzalutamide (ENZA) and mifepristone were purchased from MedChem Express LLC.

EVs Isolation From Cell Lines

To collect EVs from cell supernatants, all cell lines were cultured in EV-depleted medium. To deplete EVs from culture medium, complete medium (+10% or 20% FBS) was centrifuged overnight at 30,000 rpm followed with filtration with sterile bottle-top vacuum filters with a pore size of 0.22 µm. Cells were cultured in 3-layer flasks and cell supernatant was collected every 2–3 days, centrifuged at 1,400 rpm for 10 minutes, filtered and stored at −20°C. EVs were isolated from collected supernatant in two consecutive ultracentrifuges steps. The first step was at 30,000 rpm 4°C for 2 hours and second at 35,000 rpm 4°C for 2 hours. Pellets were resuspended in different buffers depending on the assay: for Nanosight [nanoparticle tracking analysis (NTA)]: approximately 20 mL of PBS with 2 mmol/L Ethylenediaminetetraacetic acid (EDTA); for Cell treatments: approximately 20 mL of sterile medium with 2 mmol/L EDTA; for protein: approximately 20 mL of 1X RIPA; for RNA extraction: 700 mL of QIAzol Lysis Reagent. Resuspended or lysed EVs were stored in −80°C until future experiments.

NTA

For the determination of particle concentration and size, NS500 (NanoSight Limited), equipped with an 8 mega pixel camera (Andor Technology) and a 405 nm laser, was used to measure the size and determine concentration of EVs. NTA v3.2 software (NanoSight Limited) was used for both data acquisition and analysis. The duration of each video was 5 × 1 minute. During the analysis procedure, the camera level was 14 and the detection threshold was 3. Samples were diluted from 1: 100 to 1:1,000 in PBS 2 mmol/L EDTA.

Gene Expression and Protein Expression in Cells and Derived EVs

RNA was isolated from cells and EVs using RNAeasy micro kit (Qiagen) according to manufacturer's protocols and RNA concentration was determined using a NanoPhotometer. Two-step RT-PCR was performed for gene expression analysis. In the first step, 100 ng of total RNA (concentration 10 ng/µL) from cells or EVs was used to synthesize cDNA with SuperScript IV VILO Master Mix (Thermo Fisher Scientific) according to manufacturer's protocols in a thermocycler; in the second step, 10 µL of cDNA were mixed with PowerUp SYBR Green Master Mix and specific target primers at concentration of 450 nmol/L (KiCqStart Syber green primers, Sigma) for gene expression analysis in RT-PCR instruments (Applied Biosystems QuantStudio Real-Time PCR). Standard cycling mode was running for 40 cycles according to PowerUp SYBR Green Master Mix manufacturer's protocols. Syber green primer sequences used in this project are: Human NR3C1, reverse primer sequence-GATTTTCAACCACTTCATGC-forward primer sequence -ACTGCTTCTCTCTTCAGTTC-; Human KLK3, reverse primer sequence -AGAATCCTCTGGTTCAATG- forward primer sequence – TATGAGCCTCCTGAAGAATC-. Human β-actin, reverse primer sequence -GCCCACATAGGAATCCTTCTGAC- -AGGCACCAGGGCGTGAT- was used as a control for the cells and Human 18s, -reverse primer sequence TTATCTAGAGTCACCAAGCC-, forward primer sequence -CAGTTATGGTTCCTTTGGT- as control for RNA from EVs. Cells and EVs pellets were lysed using 1 x RIPA buffer (containing 50 mmol/L Tris, 150 mmol/L NaCl, 1 mmol/L EDTA, 1% IGEPAL, 1% glycerol), supplemented with Complete Protease Inhibitor Cocktail (Roche), PhosSTOP Phosphatase Inhibitor Cocktail (Roche), 100 mmol/L vanadate (Invitrogen Life Technologies), and 1 mmol/L dithiothreitol (Sigma-Aldrich). Protein concentration was determined with Bradford assay using Bradford solution (Bio-Rad). Proteins were separated in 10% or 12% Bis-Tris gels and transferred to polyvinylidene difluoride membranes using the Trans-Blot Turbo System. Membrane was incubated with primary antibodies in the appropriate dilutions (1:500–1:1,000) overnight at 4°C. The following antibodies were used: GR from Cell Signaling (D6H2L), AR from Santa Cruz Biotechnology (441), KLK3 from cells signaling (D11E1), and GAPDH from Cell Signaling (14C10). Horseradish peroxidase–conjugated antibodies in 1:1,000 dilution was used as secondary antibodies for 1 hour at room temperature. Image manager system was used for proteins detection and analysis (KwikQuant, Kindle Biosciences, LLC)

Transfection with Short Hairpin RNA Lentiviral Particles

Mission custom short hairpin RNA (shRNA) lentiviral particles were purchased from Sigma (Sigma Millipore). The sequence of the NR3C1 shRNA-II was as follows: 5′-GTGTCACTGTTGGAGGTTATT-3′; a MISSION pLKO.1-puro non-Target shRNA Control Transduction Particles was used as a negative control. LNCaP cells were seeded at 4 × 10⁵ cells/well in a 6-well plate prior to lentiviral particles infection and incubated with 2 mL of complete medium without antibiotics for 24 hours. The day after the cells were previously coated with 8 µg/mL polybrene (Sigma Millipore) and then infected by the addition of the shRNA lentiviral particles at a multiplicity of infection of 10–20; the plates were then gently swirled for mixing and incubated for 24 hours. The virus-containing medium of infected wells was removed, and fresh complete medium was added. On day 3, puromycin selection was added at a final concentration of 5 µg/mL in all wells except the puromycin (−) control, for 3–5 days. qRT-PCR and Western blot analysis were performed to determine the shRNA interference efficiency (GR primers and GR antibody for Wester blotted reported previously).

Animal Studies

NOD/SCID mice were provided by MD Anderson Cancer Center, Experimental Radiation Oncology, Department of Veterinary Medicine & Surgery, authorized by the Jackson Laboratory. All animal studies were conducted in accordance with the current regulations and standards of the U.S. Department of Agriculture, the U.S. Department of Health and Human Services and the NIH and were approved by The University of Texas MD Anderson Cancer Center Institutional Animal Care and Use Committee. NOD/SCID mice were subcutaneously injected with LNCaP cells around 5 × 10⁶ cells/mouse mixed with Matrigel 1:1. After the tumor reached the size of 100 mm³, mice were treated with 10 mg/kg ENZA (MedChem Express LLC) diluted in 5% DMSO: 40% PEG400: 5% TWEEN-80: 50% WATER. ENZA was administrated by oral gavage 5 days/week. Mifepristone (MedChem Express LLC) was given in combination with ENZA for selected mice at the concentration of 10 mg/kg and administrated by oral gavage 5 days/week. Tumor growth was recorded twice weekly while on treatment for approximately 50 days. At the end of experiment, mice were euthanized by asphyxiation with carbon dioxide and tumor and blood were collected for future analysis.

PDX Model Studies

MDA patient-derived xenografts (PDX) were developed in the laboratory of Dr. Navone at the “Prostate Cancer Patient Derived Xenografts Program”, Department of Genitourinary Medical Oncology, MD Anderson Cancer Center, and the David H. Koch Center for Applied Research of Genitourinary Cancers. PDXs were established following previous described procedures and propagated as subcutaneous xenografts in 6 to 8 weeks old male NOD/SCID gamma male mice (Experimental Radiation Oncology, MD Anderson Cancer Center). Maintenance of PDXs in mice was approved by the Institutional Animal Care and Use Committee of The University of Texas MD Anderson Cancer Center. MDA PDX 322-2-6a (Negative GR) were implanted and treated with ENZA (MedChem Express LLC) 10 mg/kg when the tumor size reached 200 mm³. ENZA was diluted in 5% DMSO: 40% PEG400: 5% TWEEN-80: 50% WATER and administrated by oral gavage 5 days/week. Mifepristone (MedChem Express LLC) was given in combination with ENZA for selected mice at the concentration of 10 mg/kg and administrated by oral gavage 5 days/week. Tumor growth was recorded twice weekly while on treatment for around 50 days. At the end of experiment, mice were euthanized by asphyxiation with carbon dioxide and tumor and blood were collected for future analysis.

IHC Procedure for GR Antibody on Mouse and Human Tissues

We performed IHC analyses of GR expression in mouse tissue specimens obtained from MDA PDX 322-2-6a at the end of the experiment and from prostate tissue specimens obtained by prostatectomy. GR (Clone 41; Biosciences – catalog no.: 611226) was used at dilution 1/150 (mouse tissues) and 1/300 (prostate human tissues) for 1 hour at room temperature, washed and incubated with Envision Dual Link System (DAKO, catalog no.: K4061) for 30 minutes at room temperature. After that the slides were covered in 3,3ʹ-Diaminobenzidine solution (DAKO, catalog no.: K3468) and incubate for 10 minutes, and counterstained with CAT Hematoxylin (Biocare Medical, catalog no.: CATHE-G) and Bluing Reagent (Thermo Fisher Scientific, catalog no.: 22-050-114). Dehydration through 95% ethanol, 100% ethanol, was followed by clarification in xylene. At the end, coverslip were added with mounting medium. For semiquantitative/quantitative analysis, the stained slides were digitalized in an Aperio AT2 scanner (Leica Biosystems) and quantified using ImageJ software.

Total EVs Isolation From Human Plasma for Protein Isolation and NanoSight Analysis

A total of 500 mL of plasma samples was thawed in ice and mixed with 500 µL of cold PBS before centrifuge at 12,000 × g for 45 minutes at 4°C to remove cellular debris. The supernatant was transferred to an ultracentrifuge tube, and 9 mL of cold PBS was added, and samples were ultra-centrifuged for 2 hours at 35,000 rpm. After centrifuge, the EV pellet was directly resuspended in 20 µL of 1x RIPA for protein lysis. Protein concentration was determined with Bradford assay using Bradford solution (Bio-Rad). A total of 20 µL of PBS 2 mmol/L EDTA for NTA. Samples were stored at −80°C.

Total EVs Isolation and RNA Extraction From Human Plasma

To isolate EVs from patient's samples, 500 µL of human plasma were thawed on ice and centrifuge at 500 × g for 10 minutes at 4°C. For 500 mL of human plasma, 100 mL of total exosome isolation kit (Invitrogen, by Thermo Fisher Scientific) was added and homogenized by vortexing. After 30 minutes of incubation on ice, the samples were centrifuged for 10 minutes at 10,000 × g at room temperature and the EV pellet was resuspended in 100 µL of PBS. Total RNA extraction from isolated plasma-derived EVs started by adding 700 µL of Qiazol Lysis Reagent and homogenizing by vortex. miRNeasy Micro Kit (Qiagen, catalog no.: 217084) was used, according to the manufacturer's instructions. The RNA quantification was performed with a NanoPhotometer. After isolation, RNA purity and quantification analysis were performed by an RNA 6000 Pico kit on the Agilent Bioanalyzer 2100 system.

Transcriptome Analysis: Affymetrix Clariom D Assay

RNA profiling of EVs was performed by Sequencing and Microarray Facility at the University of Texas MD Anderson Cancer Center. Affymetrix's Clariom D assay provided a detailed view of the transcriptome with all known coding and noncoding splice variants. Transcriptome Analysis Console (TAC) Software was used to analyze and visualize global expression patterns of genes, exons, pathways, and alternative splicing events.

Steroid Derivatization Using Hydroxylamine Hydrochloride and LC/MS-MS Quantification

A total of 10 µL of C-13 testosterone was added as the internal standard to 500 µL of previously frozen plasma samples and kept on ice. Liquid-liquid extraction was performed using Methyl t-butyl ether (MTBE; 3.0 mL). After centrifugation, the organic layer was then collected. The sample was then derivatized using methanol (100 µL) and hydroxylamine hydrochloride (1.5 mol/L; 100 µL) heated at 65°C for 60 minutes. An 8-point standard calibration curve for the androgens was created in charcoal-stripped plasma. The samples were then transferred into LC/MS vials and analyzed (Agilent 6495 Mass Spectrometer with Jet Stream Technology and ESI source). Standard compounds and Internal Standard compounds were obtained from Sigma: testosterone (Steraloids, A6950-000), cortisol (Steraloids, Q3880-000); testosterone (³C₁₃) (Isosciences, S6066); cortisol (²H₄) (Isosciences, S7193).

Statistical Analysis

For the in vitro and in vivo studies and for the patient's samples, GraphPad Prism 9 was used for statistical analysis. Data are presented as means ± SEM. Unpaired two-tailed Student t tests were used to compare the means of two samples. Differences were considered statistically significant when P < 0.05. Gene set enrichment analysis (GSEA) was performed using the GSEA v4.3.2 software. Hallmarks and Kyoto Encyclopedia of Genes and Genomes analysis were used for enrichment of Androgen Response and Prostate cancer signaling pathway-related genes. All the gene set files were obtained from GSEA website (www.broadinstitute.org/gsea/). Enrichment map was used for visualization of the GSEA results. Enrichment score and FDR value were applied to sort Androgen Response and Prostate Cancer pathway enriched. The IPA system (version 90348151, Ingenuity Systems; Qiagen China Co., Ltd.) was used for bioinformatics analysis, which included canonical pathway analysis. The z score is generated on the basis of hypergeometric distribution and the logarithm of the significance level is obtained by right-tailed Fisher exact test.

Data Availability

All data are available in the main text or the Supplementary Materials and Methods. Further information related to the materials is available from the corresponding author.

GR is Detected in EVs of ENZA-treated Cells In Vitro

It was previously shown that GR expression was upregulated in LREX, an established ENZA-resistant cell line that expresses high levels of AR but not KLK3 (PSA) due to ASI (2). We used LREX cells to examine whether these transcripts are enriched in the EVs isolated from the conditioned medium of these cells. EVs isolated from these LREX cells, characterized by size and concentration using NTA (Fig. 1A), showed KLK3 and GR levels proportional to those expressed in the cells by qRT-PCR (Fig. 1B) and Western blot analysis (Supplementary Fig. S1A). We then treated LNCaP cells with 1 µmol/L ENZA for 5 days, to induce GR expression, as shown previously (2, 12). NTA of EVs isolated from these LNCaP cells, revealed that the size of EVs secreted by LNCaP cells were similar with or without ENZA treatment, but the EVs concentration was lower compared with control DMSO-treated cells (DMSO; Fig. 1C). RNA and proteins were extracted from cells and cell-derived EVs for qRT-PCR and Western blot analysis, respectively. In ENZA-treated LNCaP cells, GR mRNA was induced and KLK3 mRNA was decreased, as measured by qRT-PCR, compared with control LNCaP cells in both cellular RNA and RNA extracted from EVs (Fig. 1D). Similar results were obtained in protein levels as shown by Western blot analysis (Supplementary Fig. S1B). Western blot analysis of cell lysates and EVs showed different enrichment of TSG101, an EV marker that is used as loading control (Supplementary Fig. S1A and S1B). High Sensitivity RNA ScreenTape EUKARYOTIC RNA ANALYSIS was also used to demonstrate that the quality, quantity, and sizes of total RNA and RNA extracted from EVs, from LNCaP cells treated with or without ENZA, were similar (Supplementary Fig. S2A). These results suggest that GR transcript can be detected in EVs, and its level correlates with changes in the cellular GR RNA level.

To further determine whether the levels of GR transcript in EVs, reflect the changes in the cellular GR, we transduced LNCaP cells with two puromycin-inducible lentiviral shGR vectors (shGR-I, shGR-II; Supplementary Fig. S3A and S3B). Western blotting and qRT-PCR confirmed the knockdown efficiency of LNCaP cells transfect with shGR-II (GR−) compared with control LNCaP cells (CN) and LNCaP transduced with the control lentiviral vector (SCR; Fig. 2A). GR knockdown resulted in decreased cell number and cell proliferation of LNCaP cells (Fig. 2B), consistent with a previous report that GR plays a role in cancer cell proliferation (12). Knockdown of GR decreased the expression of GR in LNCaP cells as well as in the EVs by Western blot analysis and qRT-PCR compared with the EVs from control LNCaP cells (CN) and LNCaP transduced with scramble shRNA (SCR; Fig. 2C). Next, we examined whether the response of GR knockdown impacted the response of cells to ENZA treatment. As demonstrated previously, ENZA treatment led to an increased expression of GR in LNCaP CN and SCR cells but not in the GR− cells by Western blotting and qRT-PCR (Fig. 2D; Supplementary Fig. S3C). ENZA treatment reduced proliferation and increase cell death in CN, SCR and GR− cells. (Supplementary Fig. S3C and S3D).

Measuring Changes in GR mRNA Levels in Plasma EVs in LNCaP Tumors In Vivo

To examine whether changes in GR transcript in EVs could be detected in the plasma of mice bearing LNCaP tumors, LNCaP cells were inoculated subcutaneously in NOD/SCID mice (5 × 10⁶ cell/mouse). At 3 weeks postinjection (tumor size 60–100 mm³), the mice were randomized and divided into two groups, (i) VEHICLE (18 mice) or (ii) treated with ENZA (ENZA) 10 mg/kg (50 mice). The mice were treated 5 days per week by oral gavage and tumor growth was monitored weekly. Blood was drawn from these mice at baseline (BL; 200 µL which was sufficient to run qRT-PCRs), at 14 days, and at the end of treatment (END; Fig. 3). EVs were isolated from the plasma and the extracted RNA was subjected to qRT-PCR for the detection of changes in GR expression. qRT-PCR was selected as the method of choice for the detection of GR changes in the remainder of the study due to high sensitivity of the assay compared with relatively larger number of EVs required for Western blotting.

After 14 days, tumor growth showed differential kinetics in response to ENZA. We noticed that in mice treated with ENZA, there one group of animals (= 20) with tumors that decreased in size with treatment (ENZA sensitive, ENZA-S) and a second group ( = 20) with tumors increasing in size (ENZA early resistance, ENZA-R; Fig. 3A). qRT-PCR analysis showed that in mice with decreased tumor growth (ENZA-S) in response to ENZA, GR detected in the plasma-derived EVs, was significantly reduced at 14 days compared with BL (Fig. 3B). In contrast, in ENZA-R mice, GR levels were induced in EVs (Fig. 3B). The differences in GR expression in EVs were due to ENZA because they were not observed in the vehicle group at 14 days (Fig. 3B). Mice continued treatment for a total of 56 days and tumor growth size was recorded weekly (Fig. 3C). The group of mice that demonstrated decreased tumor growth (ENZA-S) had low GR expression in EVs compared with the vehicle control (Fig. 3D). The mice that demonstrated increase tumor growth (ENZA-R) had increased GR expression in EVs compared with vehicle control (Fig. 3D). We noticed that 7 of 22 mice in the ENZA-S group showed an increase in tumor volume at the end of the experiment (ENZA-R LATE), consistent with the late acquisition of resistance to ENZA. Notably, the GR levels in EVs from these 7 ENZA-R LATE mice were increased, consistent with GR-depending mechanism (Fig. 3E).

Measuring Changes in GR mRNA Levels in EVs, in MDA PCa 322-2-6a (PDX model)

Compared with cell lines, human PDXs more closely recapitulate the characteristics of the parental tumor. To further establish the feasibility and relevance of using plasma-derived EVs to monitor expression of GR and ENZA resistance, we studied a PDX MDA prostate cancer (PCa) 322-2-6a that is AR+/GR−. MDA PCa 322-2-6a tumors were subcutaneously transplanted in NOD/SCID mice and mice were randomized and divided in two groups when the tumor size reached the range of 100–300 mm³ (Fig. 4). Eight mice were treated with vehicle and 20 mice were treated with 10 mg/kg ENZA. The mice were treated 5 days per week by oral gavage and tumor growth was monitored weekly. Blood (200 µL) was drawn from these mice at BL, at 14 days and the end of treatment for exosome isolation.

Similar what was observed with LNCaP tumors, MDA PCa 322-2-6a tumors showed a differential response to ENZA treatment, with a group that responded (ENZA-S = 5 mice) and a group that did not (ENZA-R = 10 mice; Fig. 4A). qRT-PCR analysis showed that the GR mRNA levels in EVs, correlated with the tumor growth after 14 days of ENZA treatment (Fig. 4A). In the ENZA-S group, GR levels in EVs were low compared with BL, while in the ENZA-R group GR levels in EVs was significantly elevated (Fig. 4A). The ENZA-S and ENZA-R mice were treated continually with ENZA with daily oral gavage for 80 days and tumor growth was recorded weekly (Fig. 4B). Overall, ENZA treatment led to significant decrease in tumor volume compared with vehicle group. GR expression in EVs was low in ENZA-S mice and high in ENZA-R mice (Fig. 4B). These data suggest that tumor volume correlates with the plasma derived EVs GR transcript levels quantified by qRT-PCR.

Targeting of the ENZA-resistant GR+ tumors with a GR Inhibitor Leads to a Decrease in GR Levels in Tumor and Plasma-derived EVs

To investigate whether ENZA-induced GR upregulation played a role in ENZA-R tumor growth, we treated the ENZA-R tumors with mifepristone, a GR inhibitor (13–15). Mice inoculated with LNCaP or PDX MDA PCa 322-2-6a, were first treated with ENZA and after 14 days, the ENZA-R (20/42) mice were divided in two groups. One group was maintained with ENZA only treatment and the other group was treated with ENZA plus mifepristone (COMBO; Fig. 5). Treatment with COMBO started between 21 and 28 days after treatment with ENZA and at a tumor size of approximately 300 mm³. In LNCaP model, the mice that were continually treated with ENZA alone had increasing levels of GR in the plasma-derived EVs (Fig. 5A) whereas the mice that were treated with the COMBO had low levels of GR throughout the treatment. We found that in COMBO group, tumor growth was attenuated for about 2 weeks before it started growth resumed (Fig. 5B). At the end of the LNCaP experiment, that is, 56 days, GR levels in EVs in COMBO group were low compared with vehicle control group, despite tumor regrowth from COMBO treatment (Fig. 5C), suggesting that resistance to mifepristone is GR-independent. The ENZA plus mifepristone COMBO study was also examined using the MDA PCa 322-2-6a PDX model (Supplementary Fig. S4). As was observed in LNCaP tumors, induction of GR mRNA was detected in plasma-derived EVs, at 14 days posttreatment in the ENZA-R group (10 mice). The ENZA-R mice were divided in two groups, with 5 mice continued treatment with ENZA alone and 5 started on COMBO (Supplementary Fig. S4). At the end of the experiment, the COMBO group showed a longitudinal decreased in GR expression (Supplementary Fig. S4A) and an initial lower tumor growth than the ENZA alone treated group. However, the tumor growth resumed after a transient growth inhibition (Supplementary Fig. S4B). Thus, in both LNCaP and MDA PCa 322-2-6a models, inhibition of GR by mifepristone inhibited tumor growth in ENZA-resistant tumors, but the effect was transient, suggesting AR- and GR-independent resistance mechanics (Supplementary Fig. S4C).

At the end of experiment, tumors from each mouse were collected and IHC staining were performed for GR expression and quantified with Image J software (Fig. 5D). The images and relative quantification indicated that the levels of GR in EVs were low in the ENZA-S group, increased in the ENZA-R and decreased in the COMBO group at the end of experiment (END; Fig. 5D). Consistently, the intracellular levels of GR detected by IHC showed that GR protein levels were significantly reduced by the GR inhibitor mifepristone in the COMBO group indicating that the inhibitor was hitting its target (Fig. 5D). Levels of intracellular GR mRNA (TUMORS) and GR mRNA of EVs correlated with GR protein levels by IHC in the three groups of mice. The ENZA-S tumors had low expression of GR mRNA in the tissues (Fig. 5E) and in the plasma-derived EVs (Fig. 5E), the ENZA-R tumors had proportionally similar higher levels of GR in the tissues and in the plasma-derived EVs, while in the COMBO group GR was inhibited by mifepristone (Fig. 5E).

These data provide further evidence that GR mRNA levels of EVs correlate well with intratumor GR mRNA and protein levels, and this technology can be used as a pharmacodynamic measure for efficacy of targeting GR.

Measurement of GR mRNA in EVs Isolated from Liquid Biopsy of Patient Samples Reveals Differential Subset of Patients

To examine the feasibility of EVs GR detection in the clinical setting, we used patient plasma samples collected before and after 6 months ADT and ASI. In this clinical trial (NCT03279250), patients with localized high-risk or unfavorable intermediate-risk prostate cancer were treated with LHRH analog plus apalutamide (ARM A) for 6 months prior to radical prostatectomy. Plasma samples were collected before and after 6-month treatment. RNA was extracted from EVs, and total transcriptome profiling was performed by Clariom D for selected genes (Fig. 6). The PSA became undetectable or nearly undetectable in all men included in this analysis (Fig. 6A and B). GSEA showed changes in the Androgen Response pathway with a normalized enrichment score of −2.3, P value 0.007 (Supplementary Fig. S5A–S5C). Steroids levels analysis revealed a significant decrease in testosterone concentration and an increase in cortisol levels (Fig. 6C). We identified two subgroups of patients, those whose GR levels in EVs increased in response to ADT + ASI (GR+, 7 patients) and those that did not (GR−, 10 patients) as determined by the Clariom D Transcriptome Array (Fig. 6D). The GR-responsive genes PGK1, ALOX5AP, and CHCHD7 showed significant increase in expression only in the GR+ group (Supplementary Fig. S5D). To correlate the observed changes in GR expression in EVs with tumor tissue, prostatectomy samples from the same patients were stained for GR protein (Fig. 6E). Representative IHC images of 4 selected patients from the GR+ group (patient 59 and 32) and the GR− group (patient 43 and 60) demonstrated presence of GR expression in prostate cancer tumor cells, whereas lower GR expression was observed in the patients from the GR− group (Fig. 6E). Quantification of GR expression in tissues staining for all patients GR− and GR+, revealed a good correlation between expression of GR mRNA in EVs and expression of GR protein in tissue biopsies (Fig. 6E).