The experimental procedures in animals were approved by the National animal welfare and ethics committee (permit no. 121/2021) and were performed in compliance with the Directive 2010/63/EU as adopted in the national legislation.

In total, 37 naïve SPF male Fischer 344 rats were obtained from Charles River Laboratories, Germany (immunocompetent inbred strain F344/DuCrl). During the introduction, animals were randomly allocated in cages in pairs or trios; individually ventilated cages GR900, HEPA-ventilated by SmartFlow air handling unit (Tecniplast, Italy) at 75 air changes per hour were used for animal housing. Access to autoclaved water acidified to pH 2.5–3.0 with HCl and standard rodent diet (4RF21 (A), Mucedola) was provided ad libitum. Aspen wooden bedding and nesting material (Tapvei, Estonia) together with rat cardboard houses (Velaz, Czech Republic) and aspen gnawing bricks (Tapvei, Estonia) were provided in all cages, and cages were changed every 7 days. Animals were housed in SPF facility under controlled temperature (24 ± 1°C) and relative humidity of 40–60%. Animal health monitoring was performed in line with FELASA recommendations (Mähler et al., 2014).

All animals were subjected to at least 2-weeks acclimatization period with the adaptation to a reverse 12 h light/dark cycle (dark phase set to 10:00 am–10:00 pm; visible light intensity <25 lux), and then used to model either regular physical exercise (i.e. forced wheel running model) or orthotopic PC development [PLS10 rat PC model (Suzuki et al., 2015)]—see Figure 1 for schematic overview. An individual animal served as an experimental unit in both models. Before starting the procedures, the animals were identified by tattooing their tails using AIMS™ NEO-9 Neonate Tattooing System according to the manufacturer’s instructions.

Rat prostate cancer cell line PLS10 was developed previously using chemically-induced prostate carcinoma in F344 rats (Nakanishi et al., 1996; Suzuki et al., 2015). Cells were maintained in RPMI-1640 medium, supplemented with 10% FBS, 2 mM L-glutamine and 1x Antibiotic-Antimycotic at +37°C in a humidified atmosphere containing 5% CO2.

Cells in the exponential growth phase were trypsinized, counted and resuspended in PBS in aliquots of 5 × 10⁶ cells per 25 μL of PBS. The aliquots were immediately transferred to the animal facility and kept on ice water until injection. Just before injection into the rat prostate, the cell suspension was mixed with an equal volume of ice-cold Matrigel (Corning #356237, USA). Fresh cell suspensions were prepared for every batch of 3-4 animals undergoing laparotomic surgery.

In total, 16 7-week-old male F344 rats were subjected to model regular exercise and sedentary lifestyle. The number of animals per group (n = 8) was chosen based on the available data on average EV plasma concentrations and calculations we made to ensure the necessary amount of EVs for the 6-weeks injection period in the syngeneic PC model animals (Figure 1). In the second week after arrival and acclimatization to the reversed light/dark cycle, rats were introduced to the forced running wheel (system for rats, model 80805A, Lafayette Instrument) without a specific running mode. It was followed by a 12-days training phase for all 16 experimental animals (started at 9 weeks of age) by gradually increasing the running speed and duration. During the training phase, 8 best performers were selected for the runner group while the remaining 8 rats were assigned to the sedentary group. After assigning to groups, the animals were kept co-housed with their initial cage mates.

The regular forced running exercise phase for the runner group lasted for 5 weeks in total, including 4 weeks of regular running in a setting described in Figure 1. After 4 weeks of regular exercise, rats underwent brief anesthesia with 1.5% isoflurane and 500 μL of blood from the tail vein were taken. This step was repeated for the same animals immediately after 1 h of running. On the same day, a blood sample was collected from the sedentary rats following the same protocol. After 5 weeks of regular exercise, all the animals immediately after 1 h running were subjected to deep surgical anesthesia with 5% isoflurane and underwent terminal blood collection via cardiac puncture. Blood was immediately collected into S-Monovette Hematology EDTA K3 tubes (Sarstedt, Germany) or BD Vacutainer™ Blood Collection Tubes with K2EDTA (Fisher Scientific, USA), depending on the volume, and centrifuged at 2000 x g, +4°C for 15 min. Plasma was collected in fresh tubes and immediately frozen for further EV isolation.

EVs were isolated from rat plasma by size exclusion chromatography (SEC) using qEVoriginal/35 nm or qEV10/35 nm columns (IZON, USA) depending on the plasma volume. The SEC fractions were analyzed using ZetaSizer Nano ZS (Malvern Panalytical, UK) and the fractions containing EVs were concentrated using Amicon Ultra-0.5, Ultracel-3 Membrane, 3 kDa centrifugal filter units (Merck Millipore, Germany). EVs that were used for RNA sequencing were treated with proteinase K (Thermo Fisher Scientific, USA) and RNAse A (Thermo Fisher Scientific, USA) to remove all the free proteins and RNAs that are not enclosed in EVs. EVs were visualized by transmission electron microscopy (TEM) and the size distribution profile and concentration of EVs were determined by nanoparticle tracking analysis (NTA) using NanoSight NS500 instrument (Malvern, UK) as described before (Endzelins et al., 2017).

RNA was extracted from plasma EVs using miRNeasy micro kit (Qiagen, USA) according to the manufacturer’s protocol. The concentration and quality of the obtained RNA were analyzed on Agilent Bioanalyzer with Agilent RNA 6000 Pico chip (Agilent Technologies, Germany).

RNA libraries were constructed using CleanTag® Small RNA Library Prep Kit (Trilink Biotechnologies, USA), the quality and concentration of obtained libraries were analyzed on Agilent Bioanalyzer using Agilent High Sensitivity DNA chip (Agilent Technologies, Germany). The libraries were cleaned using Blue Pippin DNA Size Selection with 3% gel Blue Pippin Cassette (Sage Science, USA) setting tight target length to 140 bp thus selecting fragments with size in tight range to 140 bp (126–154 bp). The libraries were diluted as required and sequenced on Illumina NextSeq500 instrument using NextSeq 500/550 Mid Output Kit v2.5 (150 cycles) (Illumina, USA).

The obtained raw data in fastq format were analyzed using ad-hoc R script pipeline, which included: adapter trimming [cutadapt (Martin, 2011)], read mapping [bowtie2 (Langmead and Salzberg, 2012)] against RGSC rat (Rattus norvegicus) genome (version Rnor_6.0), multi-aligned reads reposition [ShortStack (Axtell, 2013)], counting [Rsubread package (Liao et al., 2019)] with RGSC (version Rnor_6.0) and miRbase (Kozomara et al., 2019) annotations. For differentially expressed gene (DEG) analysis, the reads were normalized per sample, the reads mapped to features were counted and analyzed using quasi-likelihood F-tests by edgeR (McCarthy et al., 2012) package. A subset of DEGs (adj. p < 0.05) was subjected to GO terms [GOstats (Falcon and Gentleman, 2007)] and enrichment analyses (rentrez (Winter 2017), GO.db (Carlson, 2019a), org.Rn.eg.db (Carlson, 2019b) packages).

For modeling an orthotopic PC development, 21 F344 rats were used—the sample size of 7 animals per group (see Figure 1 for study groups) was calculated by statistical power analyses using G*Power software (Buchner et al., 2021) and taking into account 80% power and α = 0.05, the published tumor size variation (Suzuki et al., 2015), expected effect size and the chosen statistical test for the result analyses.

At the age of 12 weeks, each animal underwent laparotomic surgery by using an aseptic technique under 2.5% isoflurane anesthesia, and 5 × 10⁶ syngeneic rat prostate cells PLS10 in 50 μL total volume containing 50% Matrigel were injected into the ventral lobe of the prostate. The incisions were closed with Novosyn 4/0 DS19 mid-term absorbable sutures (B.Braun, Germany) and secured with surgical adhesive. An ophthalmic gel was provided during surgery and 8 ml of warm saline was administered subcutaneously immediately after surgery. Each animal received subcutaneous Meloxicam (1 mg/kg) injection during surgery and for 3 days post-surgery. Animals were single housed in sterile cages for 3 days and then returned to their home cages. Animals were closely monitored for possible postsurgical complications; the wounds were treated with furasol solution. In total, 7 rats divided in 2 sets were implanted with PC cells per day, and animals were allocated to the study groups after stratified randomization by considering this setting (Figure 1; one animal did not survive the surgery).

Starting from day 5 post-implantation, rats from all study groups via lateral tail vein received i.v. injections of 100 μL EV solution containing 1.5 × 10¹⁰ EVs in PBS or PBS every second day as shown in Figure 1. During PC development, animals were carefully monitored for possible signs of suffering (following the IACUC Policy #012), and the tumor size was estimated by palpation. After 6 weeks of EV injection, all study animals were humanely euthanized, and terminal blood samples and tissues of interest were collected and fixed in 10% buffered formalin. Primary tumors were measured, and tumor volume was calculated using the following formula: tumor volume = (width)2 x length/2.

A nonparametric one-tailed Mann-Whitney test was used to compare the tumor volume and the number of metastases between groups of animals. Fisher’s exact test was used to compare the number of animals with/without metastases between groups of animals. p-value of ≤0.05 was considered to be significant. The statistical analyses were performed with GraphPadPrism 7 (GraphPad, USA).

EVs were isolated from rat plasma before (Pre-RUN) and immediately after forced wheel-running exercise (Post-RUN) and the yield, size and purity of EVs were assessed by TEM and NTA. TEM images revealed that the majority of particles were ranging from 30 to 160 nm in diameter and had a cup-shaped morphology that is typically observed for exosomes in TEM. However, smaller particles that possibly represent lipoprotein particles and a small number of large particles of 200–250 nm in diameter were also present in the majority of the samples analyzed and we did not observe any significant differences between Pre-RUN and Post-RUN samples (Figure 2A,B). NTA showed that the major fraction of particles was in the size range from 36 to 200 nm and the concentrations of EVs ranged from 6.5 × 10⁹ to 1.9 × 10¹¹ particles per ml of plasma (Figure 2C). However, no significant differences neither in size or concentration of EVs between Pre-RUN and Post-RUN samples were observed (Figure 2D).

To determine whether exercise affects the EV-enclosed RNA cargo, we performed RNA sequencing analysis of Pre-RUN (n = 8) and Post-RUN (n = 8) plasma EVs from the runner group and plasma EVs from the sedentary group (n = 8). Analysis of the total EV RNA by Bioanalyzer showed the dominant peaks in the range from 20 to 150 nt (data not shown). The total EV RNA was used for the RNA-seq library construction without a prior size selection. Importantly, prior to the RNA extraction, EVs were treated with Proteinase K and RNase A to remove RNAs that are attached to the surface of EVs. On average, 5.9 million raw reads were obtained per library, however, on average 2.85 million reads per library remained after the quality control, adapter trimming, and filtering out the reads that were shorter than 16 nt. The average overall alignment rate to the Rattus norvegicus genome was 51.5%. The aligned reads were counted using Rsubread package (Liao et al., 2019) and RGSC rat genome and genome annotation (version Rnor_6.0). Results showed that the majority of the reads were mapped to mRNAs (73.8%), followed by miRNAs (13.6%) and rRNAs (6.7%), while the other biotypes—lincRNAs, mitochondrial rRNAs, processed pseudogenes etc. constituted less than 2% each. However, when the reads were counted using miRBase annotation, 0.6% of the mapped reads were counted as mature miRNAs and a total of 194 different miRNAs were identified.

Next, we performed differential expression analysis using edgeR package. RNAs which expression was detected in less than 4 samples were excluded from the analysis. To assess whether the RNA cargo of EVs that are released into the circulation during the forced wheel running exercise is distinct from that in the resting state, Post-RUN EVs were compared against the Pre-RUN EVs. A total of 20 differentially expressed genes (DEGs) (adj. p < 0.05)—10 upregulated and 10 downregulated during the exercise were identified (Figure 3A; Supplementary Table S1). The top 10 DEGs are shown in Table 1. All DEGs are protein-coding genes and no differentially expressed RNAs were found in other RNA biotypes. GO term enrichment analysis of DEGs revealed enrichment of genes related to unfolded protein binding (adj. p = 0.048).

To assess the long-term effects of exercise on the RNA cargo of circulating EVs, Pre-RUN EVs were compared to the plasma EVs from the sedentary control rats. A total of 52 genes were differentially expressed (adj. p < 0.05), including 50 protein-coding genes and 2 miRNAs (Figure 3B; Supplementary Table S2). Only eleven of these RNAs had higher levels in the Pre-RUN EVs than in sedentary control EVs. The top 10 DEGs are shown in Table 2. GO term enrichment analysis showed the enrichment of genes associated with molecular functions “selenium binding” and “oxidoreductase activity, acting on peroxide as acceptor” (both adj, p = 0.025).

Next, we investigated the effects of EVs released during the forced wheel running exercise on the progression of pC. Briefly, total blood plasma EVs were isolated from F344 rats subjected to forced running wheel exercise or sedentary lifestyle, and the obtained EVs were subsequently administered intravenously in F344 rats with orthotopically injected syngeneic prostate cancer cells PLS10 (established from chemically induced, castration-resistant metastatic PC). The PLS10 PC rat tumor model was adopted based on the protocol originally published by the authors (Suzuki et al., 2015). PC-bearing rats received EV injections during 6 weeks of PC progression (representing stage 1 to stage 3/4 PC development)—REV group received Post-RUN EVs from runners, SEV group—EVs from sedentary rats, PBS group—vehicle only (Figure 1).

Six weeks after implantation, tumor masses were found in the ventral prostate of all rats. Noteworthy, no suffering was observed till the endpoint for the majority of tumor-bearing animals. However, two rats from each group were humanely euthanized before the set endpoint of 6 weeks due to signs of suffering. Postmortem analyses revealed that all the animals had tumor masses developed outside the original injection site leading to the stricture of the urethra, a common complication seen in orthotopic rodent PC models [our unpublished data (Suzuki et al., 2015)]. These animals were excluded from the data analyses.

Primary tumor volume was assessed as the primary outcome measure and compared between the groups. The mean primary tumor volumes were 4532 ± 1396 mm³ in the REV group, 7000 ± 1882 mm³ in SEV group, and 6841 ± 3291 mm³ in PBS group (Figure 4A). Statistically significant differences were seen only when comparing mean tumor volume of rats that received runner-derived EVs and those of SEV group animals (reduction of the primary tumor volume in REV group by 35%, Mann-Whitney p = 0.0159), but the difference was insignificant when compared to the PBS group, due to the small number of animals and large variation in tumor size in this group. All the primary tumors were found to have profound necrosis in the center of the tumor (Figure 4B).

Apart from tumor volumes, the location and number of macroscopic metastatic lesions were assessed. The results were highly inconsistent between individual animals and groups—metastases were found in various locations, including seminal vesicles, mesenteric, iliac and intestinal lymph nodes, and lungs (Table 3) and revealed no statistically significant differences between any of the study groups in terms of numbers or locations. However, distal lung metastases were observed only in the animals from control groups, but not rats receiving Post-RUN EVs (REV group) implying for a possible protective role of the EVs in cancer progression.