Activation of (pro)renin by (pro)renin receptor in extracellular vesicles from osteoclasts

The (pro)renin receptor (PRR) is a multifunctional integral membrane protein that serves as a component of the vacuolar H+-ATPase (V-ATPase) and also activates (pro)renin. We recently showed that full-length PRR, found as part of a V-ATPase sub-complex, is abundant in extracellular vesicles shed by osteoclasts. Here, we tested whether these extracellular vesicles stimulate (pro)renin. Extracellular vesicles isolated from the conditioned media of RAW 264.7 osteoclast-like cells or primary osteoclasts were characterized and counted by nanoparticle tracking. Immunoblotting confirmed that full-length PRR was present. Extracellular vesicles from osteoclasts dose-dependently stimulated (pro)renin activity, while extracellular vesicles from 4T1 cancer cells, in which we did not detect PRR, did not activate (pro)renin. To confirm that the ability of extracellular vesicles from osteoclasts to stimulate (pro)renin activity was due to the PRR, the “handle region peptide” from the PRR, a competitive inhibitor of PRR activity, was tested. It dose-dependently blocked the ability of extracellular vesicles to stimulate the enzymatic activity of (pro)renin. In summary, the PRR, an abundant component of extracellular vesicles shed by osteoclasts, stimulates (pro)renin activity. This represents a novel mechanism by which extracellular vesicles can function in intercellular regulation, with direct implications for bone biology.

www.nature.com/scientificreports/ PRR is a multifunctional protein 15 . While it was long known to be an associated with V-ATPase 16 , and was thought to be an accessory protein required for the assembly of V-ATPase, recent structural studies show that it is a component of the mature functioning mammalian V-ATPase 17,18 . The PRR was also identified for its ability to stimulate (pro) renin to actively cleave angiotensinogen 19 . Finally, the PRR was shown to be associated with Wnt signaling 20 . The function of PRR is associated with its cleavage by furin 21 . Intact V-ATPase isolated from mammalian brain was enriched in the transmembrane domain of the PRR 17,18 . Likewise the soluble extracellular domain released by furin has been thought to be the physiological stimulator (pro)renin 15 .
Based on the presence of full length PRR in osteoclast EVs, we hypothesized that the EVs from osteoclasts would be able to activate (pro)renin to cleave angiotensinogen. In the current study, we have tested this idea.

Results
EVs from RAW 264.7 osteoclast like cells and primary osteoclasts contain full length (pro) renin receptor and activate (pro)renin. RAW 264.7 osteoclast-like cells and primary osteoclasts (Fig. 1A,B), were cultured and EVs were isolated. EVs were characterized by nanoparticle tracking (Fig. 1C,D) The sizes of EVs from RAW 264.7 osteoclast-like cells and primary osteoclasts were in the range of extracellular vesicles, but EVs from osteoclasts were smaller on average. This is consistent with the small size of EVs we have detected previously from primary osteoclasts.
Immunoblots for PRR and RANK confirmed their presence in EVs from RAW 264.7 osteoclast-like cells and primary osteoclasts (Fig. 2). As shown previously 13,14 , EVs from osteoclasts and osteoclast-like cells contained similar amounts of RANK. PRR was also detected in similar amounts and was 38 kD, consistent with the full-length protein. The tetraspanin CD81 was present as a positive control for EVs, while failure to detect the endoplasmic reticulum protein GP96 supports that there was little non-EV contamination. GP96 was detected in whole cell extracts of osteoclasts, while the PRR was not detected in EVs isolated from 4T1 cells (Fig. 2).
EVs from osteoclasts and osteoclast-like cells dose-dependently stimulated (pro)renin to become proteolytically-active (Fig. 3A,B). Controls were EVs from 4T1 breast cancer cells, in which we were unable to detect PRR. EVs from primary mouse osteoclasts also dose-dependently stimulated (pro)renin activity (Fig. 3C,D). Controls included EVs from 4T1 cells and no EVs.
The "handle region peptide" from (pro)renin blocks the ability of PRR in EVs to activate (pro) renin to cleave its substrate. The "handle region peptide" contains the sequence from a binding site in (pro)renin for the PRR, and has been shown to be a competitive inhibitor, blocking PRR's ability to stimulate (pro)renin 22,23 . The handle region peptide dose-dependently blocked the ability of PRR-containing EVs to stimulate (pro)renin to be enzymatically active (Fig. 4A,B).

Discussion
In this study, we identify a new potential regulatory mechanism associated with EVs, the ability of EVs that contain PRR to stimulate (pro)renin to become enzymatically active. Osteoclasts shed large amounts of PRR in EVs 11 , and stimulation of RAS is bone catabolic 7,8 . Both osteoclasts and osteoblasts have AT2 receptors and respond to RAS 24 . These data suggest that by stimulating RAS, EVs shed by osteoclasts may contribute to the stimulation of osteoclast activity directly (autocrine), or indirectly through osteoblasts (paracrine) (Fig. 5).
Three physiologic roles have been identified for the PRR: accessory protein/component of V-ATPase 16-18 , stimulator of (pro)renin 25,26 , and as a scaffold between V-ATPase and Frizzled, which stimulates Wnt/beta catenin signaling 20 . PRR is normally not present on the cell surface of most cells, and the ability to stimulate (pro)renin has been considered primarily a function of the cleaved extracellular domain. The fact that full length PRR, with the ability to stimulate (pro)renin, is present in EVs provides another route by which PRR can be exposed to the extracellular milieu, and potentially activate RAS.
The intact V-ATPase is composed of 16 subunits 17,18,27 . We reported that the PRR in osteoclast EVs was associated with a sub-complex of the V-ATPase 11 . This was intriguing since it is known that the PRR is required for V-ATPase assembly, although its exact role and mechanism in assembly are not known 15 . Very recent structural data show that membrane domains of PRR, along with a second accessory protein ATP6AP1/Ac45, are fully integrated into the functional mammalian V-ATPase from rat and bovine brains 17,18 . It is possible that the subcomplex found in osteoclast EVs represents a V-ATPase assembly intermediate that when exported in EVs gains a "moonlighting" function as a stimulator of RAS.
We used EVs from 4T1 breast cancer cells as a PRR-deficient control. We had originally expected that these cells, and their EVs, might be enriched in PRR and other V-ATPase subunits because plasma membrane V-ATPase has been shown to be involved in their metastatic activity 28 . However, when directly compared with osteoclasts the level of V-ATPase, and PRR is much lower in 4T1 cells compared with osteoclasts.
In summary, osteoclasts shed EVs that contain large amounts of uncleaved PRR. The PRR in the EVs stimulates (pro)renin to become active and thus would be expected to trigger increased local RAS, which favors bone resorption. Future studies will be necessary to confirm this idea. The ability to stimulate (pro)renin represents a new potential mechanism by which EVs can serve as intercellular regulators, and possibly, a new mechanism for controlling bone remodeling.  www.nature.com/scientificreports/ Cell culture. Primary osteoclasts were grown from precursors obtained from mouse femora and tibiae as described 13 . The University of Florida Institutional Animal Care and Usage Committee approved all mouse protocols (University of Florida IACUC protocol number: 20180303097). The study adhered to the ARRIVE guidelines. All procedures were performed in accordance with guidelines of the National Institutes of Health of the United States. Mice (C57BL/6, Charles River) were sacrificed by cervical dislocation, bones were dissected, and marrow was flushed from the marrow space with α-MEM complete media (Sigma-Aldrich) plus 10% exosomefree fetal bovine serum (System Biosciences) with 1% L-glutamine (Thermo Fisher Scientific), and 1% penicillin/ streptomycin/am using a syringe with a 25-gauge needle. Cells were seeded in T75 flasks at a concentration of 1.5 × 10 6 cells/mL supplemented with 5 ng/mL recombinant murine Macrophage-Colony Stimulating Factor [CSF-1] (Peprotech, Rocky Hill, NJ, USA) and allowed to grow for 24 h at 37 °C and 5% CO2. Nonadherent cells were removed, and 5.9 × 10 5 cells/mL adherent cells were seeded in 24-well plates or at 2.1 × 10 6 on 6-well plates. All cultures were supplemented with 10 ng/mL CSF-1 and 5 ng/mL soluble recombinant RANKL (sRANKL) 29 to stimulate differentiation of osteoclasts. To generate osteoclast-enriched cultures, cells were cultured for 5 or 6 days with α-MEM with 10% exosome free fetal bovine serum (System Biosciences) refreshed every 3 days. 4T1 murine breast cancer cells (kind gift of Gary Sahagian, Tufts University, Boston, MA) were grown in dMEM plus 10% exosome-free fetal bovine serum (Systems Biosciences) 30 . Conditioned media was collected while cells were 50-80% confluent.

Methods
EV isolation. All steps in EV isolations were done under sterile conditions as described previously 11,13 . Exo-Quick TC material from System Biosciences was used to isolate EVs from cultures of primary cells following the manufacturer's instructions. The final pellet, containing EVs and ExoQuick, was diluted fivefold with phosphatebuffered saline (PBS) to induce the ExoQuick material to return to the soluble state. The samples were then spun at 200,000×g for 2 h in an Airfuge (Beckman Coulter, Brea, CA, USA) and the pellets were collected.
Microscopy. Primary osteoclasts and RAW 264.7 osteoclast-like cells were fixed with 2% formaldehyde in PBS for 20 min, permeabilized with 0.5% Triton X-100 in PBS, then stained for tartrate resistant acid phosphatase (TRAP) activity using Leukocyte acid phosphatase kit (Sigma-Aldrich catalog #386A). Images were taken using a Nikon Diaphot phase contrast microscope (ELWD 0.3) at a magnification of 100 X. EVs with an anti-CD81 antibody. CD81 is a marker for EVs. The fifth panel shows an effort to stain 5 × 10 8 EVs with an anti-GP96 antibody, which supports that the EVs are not contaminated with material from the endoplasmic reticulum, a common contaminant. A whole cell extract lane from osteoclasts is presented to show that the anti-GP96 antibody works. The sixth panel shows that 4T1 EVs (1 × 10 8 ) have the expected protein marker EpCAM. The seventh panel demonstrates that we were unable to detect PRR in 1 × 10 8 EVs isolated from conditioned media of 4T1 cells. As previously reported, these data show EVs from mouse osteoclasts and osteoclast-like cells contain RANK and full length PRR. They demonstrate that the amount of PRR in EVs from 4T1 cells is much lower. www.nature.com/scientificreports/ Nanoparticle tracking. The diameter size and concentration of EV population was determined using a NanoSight NS-300 (Malvern) as described previously 29 . Samples were evaluated using different dilutions in sterile-filtered PBS and videos recording for 60 s were used to estimate the concentration and size distribution of EVs by light scattering and Brownian motion. The Nanosight NTA Software analyzed raw data videos by triplicate.
Western Blots. Protein samples were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis on 4-20% gels using the Mini-Protean system (BioRad). Gels were blotted to nitrocellulose or to Immobilon membranes (Pierce) and horse radish peroxidase-conjugated secondary antibodies were used to detect primary antibodies. These were detected either using a chromogenic substrate (ThermoFisher, CN/DAB substrate), or chemiluminescent substrate (ThermoFisher, Super Signal West Pico). Blots were either photographed or detected using a BioRad ChemiDoc MP (BioRad). The Optimal Autoexposure setting was used to acquire images. The Raw photographs and chemiluminescent data (included in Supplementary Fig. 1) were minimally (Pro)renin activity assay. The SensoLyte 520 Mouse Renin Assay kit (Anaspec, Fremont, California) was used to assay for renin activity. This kit allows for continuous assay of renin activity using a 5-FAM/QXL 520 fluorescence resonance energy transfer (FRET) peptide. In the FRET peptide, the fluorescence of 5-FAM is quenched by QXL 520. Upon cleavage into two separate fragments by mouse renin, the fluorescence of 5-FAM is recovered. This was measured with a fluorescence multiwell plate reader (Synergy HTX, Biotek, Winooski, Vermont) at excitation/emission: 〖"λ " 〗_ex = 485 nm / 〖"λ " 〗_em = 528 nm. Fluorescence readings are expressed in relative fluorescence units (RFU). Instruments were calibrated using a 5-FAM fluorescence reference standard. The plates were placed in the multiwell plate reader for 60 min, with measurements taken every 5 min.
Statistics. The results are expressed as mean plus/minus Standard Error. We used the program Graph-Pad Prism 5 (GraphPad Software, La Jolla, CA) to compare samples by One-Way ANOVA and Student's t-test. P values < 0.05 were considered significant. Nanoparticle tracking was analyzed by ANOVA with Tukey's modification 30 .

Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.