Most, if not all, cell types release membrane-bound vesicles, which can be isolated from biologic fluids such as blood,1 saliva,2 and urine.3 These extracellular vesicles (ECVs) are heterogeneous but can be broadly classed as being derived directly from the cell membrane or originating from intracellular pathways. Cell membrane–derived ECVs include microparticles and apoptotic bodies. Classically, exosomes are distinguished from other ECV subpopulations by their unique biogenesis in the intracellular endosomal pathway following fusion of the multivesicular body with the limiting plasma membrane. These intracellular vesicles are termed ‘exosomes’ after release into the extracellular space.
Conventionally, ECVs were viewed as a removal system for senescent or excess lipid and protein from cells.4 Their cargo changes with cell injury and disease, so ECVs represent a potential reservoir for biomarker discovery, even a noninvasive replacement for tissue biopsy.5 However, this promise has not yet been translated into a biomarker with real-life clinical utility. Research has identified a potential role in intercellular signaling; ECVs can deliver functional protein and RNA from one cell to another in vitro.6 The mechanisms by which target cells internalize ECVs are yet to be fully elucidated, and whether ECV transfer between cells occurs in vivo is still to be unequivocally confirmed. In cell culture studies, ECV uptake by cells has been reported via several mechanisms, including clathrin-dependent endocytosis, caveolae-dependent endocytosis, phagocytosis, and macropinocytosis.7 However, it is not established whether ECV uptake by recipient cells is a physiologically regulated process and, if it is, which pathways or hormones are involved.
Urine contains ECVs originating from the circulation and from cells that line the urinary tract.3 These ECVs maintain urine sterility by virtue of their antibacterial activity.8 The most-studied urinary ECV subtype are exosomes, which are derived from the kidney’s glomerulus and all regions of the nephron. Urinary exosomes contain protein, mRNA, microRNA (miRNA), and mitochondrial DNA that originate from kidney tubular cells.3,9 Given the unidirectional flow of urine along the nephron, the kidney is anatomically designed for potential ECV transfer from proximal to distal nephron segments. In the kidney, there is evidence of ECV signaling: Exosomes from injured tubular cells transfer mRNA into fibroblasts, resulting in cell activation, and stem cell–derived exosomes protect against AKI by transfer of RNA.10,11 Using aquaporin (AQP) membrane water channels to track ECV signaling, we previously demonstrated ECV-mediated AQP transfer from stimulated to unstimulated collecting duct cells.12
Using a kidney collecting duct cell line (mCCDC11), which responds robustly to aldosterone and vasopressin stimulation by transporting sodium and water, respectively, we have established and characterized a model of ECV release. These ECVs are exosome-like in their properties: They express archetypal exosomal proteins, such as tumor susceptibility gene 101 (TSG101), which localize to a density on a sucrose gradient that is characteristic of exosomes.13 Transmission electron microscopy and nanoparticle tracking analysis (NTA) confirm that this collecting duct cell line releases ECVs of exosome size and shape that express surface exosomal markers.12,14 We have previously demonstrated that vasopressin, a pituitary neuropeptide that regulates water homeostasis, modulates the aquaporin 2 (AQP2) content of these ECVs in vitro and this regulation translates into rodent models and humans.12,14 The aim of the present study was to investigate the role of vasopressin in the regulation of ECV uptake into the kidney collecting duct.
Results
Vasopressin Stimulates ECV Uptake by mCCDC11 and Primary Collecting Duct Cells
ECVs of exosomal size were present in the supernatant from the mCCDC11 cells, as we have previously reported (Supplemental Figure 1). Release of ECVs from mCCDC11 cells was significantly increased after stimulation with the vasopressin analogue, desmopressin (Supplemental Figure 2). When the size distribution of ECVs was analyzed by NTA, the increase in ECV release induced by desmopressin corresponded to release of ECVs in the size range 20–100 nm, an exosomal size distribution. NTA is a light-scatter microscopy method of tracking microparticles and nanoparticles on the basis of direct and real-time tracking of the particles’ Brownian movement, which results in a description of the particle size and concentration distribution in a given solution. NTA can be used to count and measure specific subgroups of nanoparticles using fluorescent antibodies against surface proteins, including ECVs derived from kidney cells in culture and in urine.14 The mCCDC11 cell ECVs were successfully loaded with fluorescent Cell Tracker (Invitrogen, Carlsbad, CA) label and membrane disruption with QIAzol cell lysis reagent substantially reduced the NTA signal, consistent with fluorescent loading of membrane-bound ECVs (Supplemental Figure 3). Cell Tracker nanocrystals in PBS without any ECVs produced no NTA signal.
ECV uptake by mCCDC11 cells was quantified by tracking fluorescence and by the cellular response to miRNA-loaded ECVs, two complementary approaches that report physical uptake and biologic activity, respectively. ECVs, harvested from mCCDC11 cells and loaded with fluorescent nanocrystals (Cell Tracker label), were applied to different mCCDC11 cells grown in a confluent monolayer. Entry of ECVs into the recipient cell was observed under fluorescent microscopy (Figure 1), the percentage of fluorescent cells was determined by FACS analysis (Becton Dickinson, San Diego, CA), and the concentration of ECVs remaining in the supernatant was measured by NTA. Incubation of the recipient cells with the vasopressin analogue, desmopressin, caused a time-dependent increase in the proportion of fluorescent cells (Figures 1 and 2, A and B). Significant ECV uptake occurred after 48 hours of desmopressin stimulation; shorter exposure times had no effect (Figure 2A). This is similar to the time course of desmopressin-induced ECV release (Supplemental Figure 2) and AQP2 expression in this cell line.12 At concentrations similar to the physiologic concentration of vasopressin,15 ninety-six hours of desmopressin incubation approximately doubled the proportion of recipient cells taking up fluorescent ECVs (Figure 2C). Consistent with this, the concentration of fluorescently loaded ECVs remaining in the supernatant was significantly reduced by treatment with desmopressin (Figure 2D). Tolvaptan, a selective V2 receptor antagonist, abolished the increase in ECV uptake induced by desmopressin (Figure 2E), while OPC-21268, a V1 antagonist, had no effect (23.2%±5.2% for desmopressin-stimulated mCCDC11 cells versus 18.8%±1.4% for desmopressin-stimulated cells treated with OPC-21268 [1 μM]; n=3; P=0.5). Endothelin-1, a peptide that inhibits sodium transport in the collecting duct,16 had no effect on ECV uptake by mCCDC11 cells when applied alone but physiologically antagonized the effect of desmopressin (Figure 2F).
;)
Figure 1.: Desmopressin stimulation causes ECV uptake. Fluorescent microscopy of unstimulated (left) and desmopressin-stimulated (3.16 ng/ml for 96 hours, right) mCCDC11 cells incubated with fluorescently loaded ECVs (red). Desmopressin stimulated ECV uptake into the cellular cytoplasm. Nuclei were stained with DAPI (blue), and cell membrane was stained with phalloidin (green).
Figure 2.: ECV uptake by mCCDC11 cells is increased by desmopressin stimulation. (A) Flow cytometry data demonstrating no significant fluorescently loaded ECV uptake following desmopressin stimulation (3.16 ng/ml) for up to 8 hours (n=3). Fluorescent cells expressed as percentage of total cell number. (B) Flow cytometry data demonstrating ECV uptake following longer desmopressin stimulation. n=3; *P<0.05: 3.16 ng/ml desmopressin stimulation versus no stimulation. For (A) and (B) error bars represent SD. (C) Desmopressin (3.16 ng/ml for 96 hours)-stimulated mCCDC11 cells had significantly increased fluorescence after incubation with labeled ECVs. Fluorescent cells expressed as percentage of total cell number (n=5). (D) NTA analyses of fluorescently loaded ECVs in the cell culture supernatant from control and desmopressin (3.16 ng/ml for 96 hours) stimulated cells presented as the area under the concentration curve for particle sizes between 20–100 nm (n=5). Desmopressin stimulation reduced the concentration of ECVs in the supernatant. (E) Fluorescence of control and desmopressin stimulated cells (3.16 ng/ml for 96 hours) exposed to fluorescently loaded ECVs in the absence and presence of tolvaptan (10 nM). Fluorescent cells expressed as percentage of total cell number (n=5). (F) Fluorescence of control and desmopressin-stimulated cells (3.16 ng/ml for 96 hours) exposed to fluorescently loaded ECVs in the absence and presence of endothelin-1 (10 pg/ml). Fluorescent cells expressed as percentage of total cell number (n=5). (C–F) Tukey plots. Bottom and top of the box are the first and third quartiles, and the band inside is the median. The whiskers are the lowest datum still within 1.5 interquartile range of the lower quartile and the highest datum still within 1.5 interquartile range of the upper quartile. *P<0.05. AUC, area under the curve.
To complement and confirm the data from fluorescence-based ECV tracking, we used ECVs loaded with a specific miRNA and used cellular target mRNA suppression as the readout of functional ECV uptake. ECVs were harvested from a human umbilical vein endothelial cell (HUVEC) line transduced to overexpress miRNA-503 (miR-503)17 and a control cell HUVEC line. This cell line was chosen because the mRNA targets of miR-503 are well described.17,18 The expression of miR-503 in the isolated ECV pellet was confirmed by quantitative PCR (overexpressing cycle threshold value of 26.6 versus control cycle threshold value >35). ECVs from both cell lines were added to unstimulated or desmopressin-stimulated mCCDC11 cells. Target genes influenced by miR-503 have been identified: We measured the mRNA expression of vascular endothelial growth factor A, fibroblast growth factor 2, and cyclin-1E and cell division cycle 25A. ECVs containing miR-503 induced a significant downregulation of the target genes after desmopressin stimulation (Figure 3). In the absence of desmopressin, there was no significant change in gene expression.
Figure 3.: ECVs deliver functional miRNA into desmopressin-stimulated mCCDC11 cells. Change in gene expression when desmopressin stimulation (3.16 ng/ml for 96 hours) is compared with unstimulated cells in the absence of ECVs, with control HUVEC-derived ECVs or with miR-503 loaded ECVs. (A) Vascular endothelial growth factor A. (B) Fibroblast growth factor 2. (C) Cyclin E-1. (D) Cell division cycle 25A. Values are expressed as the difference between control values in unstimulated cells minus control values in desmopressin-stimulated cells, with 18S as endogenous control. Negative values indicate downregulation and positive values indicate upregulation of target genes by desmopressin without or with the stated ECVs. n=9; *P<0.05. All plots are Tukey plots.
Building on these data from our collecting duct cell line, ECV uptake was visualized by fluorescence microscopy in rat primary collecting duct cells to determine whether ECV uptake was also regulated by the vasopressin system in native cells. We isolated cells expressing AQP2 from whole rat kidney (Figure 4). There was uptake of ECVs into these cells, but only in the presence of desmopressin (3.16 ng/ml for 2 hours) (Figure 4).
Figure 4.: Rat primary collecting duct cells take up ECVs after desmopressin stimulation. The cells express AQP2 protein (green), as detected by fluorescent antibody labeling. Fluorescently loaded ECVs (red) enter primary cells with desmopressin (3.16 ng/ml for 2 hours) treatment. DAPI-stained nuclei are blue.
cAMP and Dynamin Mediate Desmopressin-Induced ECV Uptake
The V2 receptor is coupled with Gs proteins and causes activation of the cAMP pathway.15 Inhibition of cAMP-dependent protein kinase A (PKA) with H-89 prevented the increase in uptake of fluorescent ECVs after desmopressin stimulation (Figure 5A). Stimulation of mCCDC11 cells with forskolin increased uptake of fluorescent ECVs independent of desmopressin stimulation (Figure 5B). Endocytosis can be cAMP dependent.19 On the basis of previous studies showing that ECVs enter cells through the endocytotic pathway,7 we investigated the role of dynamin, a guanosine triphosphatase that mediates endocytosis. Dynasore, a noncompetitive inhibitor of dynamin activity,20,21 significantly reduced desmopressin-stimulated ECV uptake to a level below that of control cells (Figure 5C). In combination, these data indicate that basal and desmopressin-induced uptake of ECVs requires cAMP activation and dynamin activity.
Figure 5.: ECV uptake by mCCDC11 cells following desmopressin stimulation is mediated by cAMP and dynamin activity. (A) Desmopressin (3.16 ng/ml for 96 hours) stimulated ECV uptake, which was decreased by PKA inhibition (H-89–25 µM) of the cAMP pathway. (B) ECV uptake was increased by cAMP stimulation through forskolin (10 µM). (C) Dynasore (150 nM) inhibition of dynamin activity decreased ECV uptake by desmopressin-stimulated cells below that of the control cells. (A–C) Fluorescent cells expressed as percentage of total cell number. n=6; *P<0.05. All plots are Tukey plots.
ECV Uptake by mCCDC11 Cells Is Selective for ECV Cell of Origin
Next, we determined whether ECVs derived from different cell types are also internalized under vasopressin regulation. We incubated mCCDC11 cells (without and with desmopressin stimulation) with equal numbers of ECVs (1×108/ml) harvested from the following renal cell types: mCCDC11 cells (mouse), proximal tubule (HK2, human), and juxtaglomerular (RG1, mouse). The ECVs from all cell types had similar size distributions, as quantified by NTA (Figure 6A). Treating recipient mCCDC11 cells with desmopressin increased uptake of the proximal tubule and collecting duct–derived ECVs, but not of those from juxtaglomerular cells (Figure 6B). This tubular cell selectivity was confirmed by NTA analysis, which demonstrated decreased proximal and collecting duct ECVs in the mCCDC11 cell culture supernatant but no change in ECVs from juxtaglomerular cells (Figure 6C).
Figure 6.: ECV uptake is cell type specific. (A) NTA measurement of ECVs from different cell types before incubation with mCCDC11 cells. (B) Comparing total cell fluorescence between control and desmopressin (3.16 ng/ml for 96 hours) stimulated mCCDC11 cells after labeled ECV incubation from three cell types: mCCDC11 cells (CCD), HK2 (human proximal tubular cell line), and RG1 (murine juxtaglomerular cell line). ECVs, 1×108/ml, were added to all experiments. n=6; *P<0.05. (C) NTA analyses of cell culture supernatant incubated with different cell type derived ECVs (mCCDC11, HK2, and RG1) from control and desmopressin (3.16 ng/ml for 96 hours) stimulated cells presented as the area under the concentration curve (AUC) for particles sized between 20 and 100 nm. Desmopressin stimulation reduced the concentration of mCCDC11 and HK2 ECVs in the supernatant but not in cells incubated with RG1 ECVs (D) Polarized cells take up ECVs under desmopressin regulation. Total cell fluorescence of polarized mCCDC11 cells stimulated with desmopressin (3.16 ng/ml for 96 hours) either apically or basolaterally compared with unstimulated cells. Labeled ECVs were applied to the apical compartment of Transwell. n=3; *P<0.05. (B–D) Tukey plots. CCD, collecting duct; HK2, proximal tubule; RG1, juxtaglomerular cell derived.
The V2 receptor is expressed on the basolateral membrane of the renal principal cell. Therefore, we examined ECV uptake in mCCDC11 cells polarized by growing on transwell plates and stimulated with desmopressin on the apical or basolateral side. Basolateral desmopressin stimulated uptake of apically applied ECVs, whereas desmopressin applied to the apical membrane had no effect (Figure 6D).
Vasopressin Regulates Urinary ECV Excretion in Mice and Humans
As a control, mice were intravenously (IV) injected with ECV-free Cell Tracker nanocrystals in solution, and no signal was detected in their urine by NTA. Mice were IV injected with fluorescently loaded ECVs derived from mCCDC11 cells. The urinary excretion of these fluorescent ECVs was measured by NTA. QIAzol treatment of the urine substantially reduced the number of particles measured by NTA, which is consistent with the presence of fluorescently loaded membrane bound ECVs (Supplemental Figure 4).
Mice were then administered tolvaptan, furosemide, or vehicle (control), followed by a second injection of the same number of ECVs. Urine output increased 1.3-fold after administration of the tolvaptan and furosemide dose. In the control group, 2.5%±1.0% and 1.8%±0.7% of the systemically injected ECVs were recovered in the urine after the first and second ECV doses, respectively. In a separate group of mice, tolvaptan treatment increased the ECV excretion from 1.1%±0.7% to 13.5%±3.9%, consistent with inhibition of vasopressin-mediated ECV uptake (Figure 7A). Treatment with furosemide produced no change in ECV excretion.
Figure 7.: Vasopressin V2 receptor regulates urinary ECV excretion in mice and humans. (A) Urinary excretion of systemically adminstered ECVs in mice. Group 1: control group; urine ECV excretion following two IV injections of fluorescently loaded ECVs. Group 2: urine ECV excretion after two IV injections of labeled ECVs. Between injections, mice were treated with tolvaptan (0.3 mg/kg). n=5 per group; *P<0.05. Group 3: urine ECV excretion after two IV injections of labeled ECVs. Between injections, mice were treated with furosemide (1 mg/kg, n=4). ECV excretion expressed as percentage of the total number of injected ECVs excreted in the urine. Data expressed as a Tukey plot. (B) Twenty-four hour ECV excretion by a patient with central diabetes insipidus. Lines represent the time of desmopressin treatment (dashed line) and concentration of ECVs expressing nephron segment–specific proteins: glomerular (podocalyxin-like protein, PODX-L), proximal tubular (cubilin), and CD24 (pan-segment urinary ECV marker). ECV urine concentration measured by NTA and normalized by urinary creatinine concentration. All data are expressed as mean±SD. NTA measurements were taken in triplicate for each time point. AUC, area under the concentration curve.
Kidney, liver, and spleen tissue was examined by fluorescence microscopy after systemic injection of fluorescently loaded ECVs without and with tolvaptan pretreatment (Figure 8). In the absence of tolvaptan, ECVs were clearly present in kidney tissue (and liver and spleen). After tolvaptan treatment, ECV numbers in the kidney decreased substantially, without an obvious effect on ECV deposition in liver and spleen (Figure 8).
Figure 8.: Tolvaptan reduces kidney uptake of ECVs. Tissue distribution of fluorescently loaded ECVs systemically injected into mice without and with tolvaptan pretreatment. Mouse organs were harvested 1 hour after IV injection of ECVs. Red, ECVs; blue, DAPI; green, AQP2. Without tolvaptan, red signal is present in kidney tissue. With tolvaptan (0.3 mg/kg) pretreatment, the red signal is reduced. ECVs are present in liver and spleen without and with tolvaptan. Bar, 20 μm.
As an additional in vivo proof-of-concept study, the urinary excretion of nephron segment–specific ECVs was measured in a patient with central diabetes insipidus using NTA combined with Qdot-labeled antibodies (Invitrogen) for segment-specific proteins.14 After self-directed desmopressin intranasal administration, both glomerular (podocalyxin-like) and proximal tubular (cubilin) protein-expressing ECVs decreased (Figure 7B). This is consistent with our cell and mouse data and supports the presence of desmopressin-regulated uptake of glomerular and proximal tubule ECVs by distal segments of the nephron in humans.
Discussion
ECVs transfer protein and RNA between cells to alter the phenotype of the recipient cell in vitro. Our data demonstrate that ECV uptake by recipient cells is hormonally regulated by vasopressin in kidney cell culture, mice, and humans. We also demonstrate that ECVs can, by vasopressin-regulated transfer of miRNA, change gene expression profiles in the recipient mCCDC11 cell. The renal tubular system is ideally suited to exploit this mechanism of cell-to-cell communication, which opens up a new paradigm for physiologic crosstalk between nephron segments. Moreover, ECVs offer a vehicle for targeting RNA therapies to diseased kidney tubular cells.
Vasopressin is released from the posterior pituitary in response to an elevation in blood osmolality. Its principal role is to stimulate water reabsorption by the renal collecting duct. This is achieved through activation of the V2 receptor on the basolateral membrane of renal principal cells, which, via a cAMP/PKA cascade, phosphorylates the water channel AQP2, thereby permitting trafficking to the apical cell membrane from subapical recycling endosomes. In parallel, vasopressin stimulates endocytosis of vesicles from the cellular plasma membrane to maintain membrane equilibrium. In this report, we demonstrate a new role for vasopressin as a hormonal regulator of ECV uptake in cells, mice, and humans. In cell culture studies, we used four complementary read-outs of ECV uptake: fluorescent microscopy, flow cytometry, and miRNA transfer into cells, combined with NTA of ECVs remaining in the culture medium. The data generated by these different methods consistently demonstrated that desmopressin stimulated ECV uptake into mCCDC11 cells and primary collecting duct cells. The time course of ECV uptake was different: mCCDC11 cells needed 48 hours of desmopressin stimulation before significant uptake was recorded, similar to the time-frame required before desmopressin stimulates AQP2 expression and ECV release in these cells. In contrast, primary cells responded within 2 hours of exposure. One explanation for the different response times is that naive mCCDC11 cells require new protein production for uptake of ECVs, but this remains speculative.
The uptake mechanism was V2 receptor mediated and cAMP/PKA dependent, in keeping with the established physiologic pathway that increases water uptake. In our cell model, desmopressin-induced ECV uptake was reduced by endothelin-1, suggesting that ECV uptake is under opposing physiologic regulation by vasopressin and endothelin-1. This is consistent with data that indicate endothelin-1, via the endothelin B receptor, antagonizes the physiologic actions of vasopressin in the collecting duct both in vitro22 and in vivo.16 Thus, the mechanism of vasopressin-induced ECV uptake is consistent with the known physiology of this hormone and is likely to be a consequence of hormone-induced plasma membrane endocytosis. Hormonal regulation of ECV entry into cells has not been demonstrated in any cell line, and we also found that vasopressin stimulated ECV (likely exosome) release from mCCDC11 cells. It is not yet clear how this uptake and release fits into a cell signaling paradigm, and it is even possible that ECVs taken up from the tubular lumen could be rereleased. Nevertheless, that both processes were influenced by vasopressin indicates that ECV intercellular signaling may be a tightly regulated physiologic process.
Publications on the biology of ECVs have substantially increased, particularly those relating to their signaling potential. However, in the kidney and other organs, there is little evidence that regulated signaling occurs in vivo. To test whether vasopressin is important for renal ECV uptake and excretion in vivo, we injected fluorescently loaded ECVs systemically into mice. After injection, these ECVs appeared in urine and kidney tissue, which is consistent with previous published studies23 and important for two reasons. First, investigators performing proteomic and transcriptomic analysis of urinary ECVs cannot assume that new biomarkers have originated from the kidney, and urinary nonrenal ECVs may offer a noninvasive way to assess the physiology and pathology on other (nonrenal) organs. Second, the presence of plasma-derived ECVs in urine provides proof of concept that systemically administered novel therapeutic interventions, delivered within ECVs, could gain access to renal tubules. The mechanism of systemic ECV entry into urine remains to be determined and a greater understanding may allow ECV manipulation to increase their urinary excretion.
We also demonstrated that tolvaptan, a selective V2 receptor antagonist, substantially increased the urinary excretion of systemically administered ECVs and reduced the number of ECVs present in kidney tissue. Furosemide had no effect despite increasing urine output, which strongly suggests that the increased excretion of ECVs after tolvaptan treatment was not due to the increased urine output per se. To our knowledge, these are the first data demonstrating that vasopressin is a regulator of urinary ECV excretion, and they confirm that our findings in cells translate to mice. Combining antibodies to nephron segment–specific proteins with NTA can identify the cellular origin of urinary ECVs. We collected urine from a patient with central diabetes insipidus (a condition defined by lack of vasopressin) and determined the effect of intranasal desmopressin on glomerular and proximal tubule–derived ECVs. After administration of desmopressin, the urinary concentration of these ECVs fell, which is consistent with vasopressin regulation of urinary ECV excretion in humans.
Although these human data are hypothesis generating, they are entirely consistent with the data from cells and mice. A limitation is that the concentration of urinary creatinine changed as a result of desmopressin treatment, making the normalization of spot urine ECV numbers a challenge. In the future, larger validation studies should be performed to confirm our human data. We conclude that kidney collecting duct cells, under the control of vasopressin, actively modify the population of ECVs in urine. This is an important concept and must be accounted for in preclinical and clinical analysis of urinary ECV profiles.
Urine contains ECVs from multiple cell types outside and within the urinary tract. The collecting duct, at the distal end of the tubular urine flow, is therefore exposed to ECVs from multiple cell types, which express a wide range of surface proteins. Intriguingly, uptake by our mCCDC11 cell line was selective with regard to the ECV cell of origin. Proximal tubular ECVs (HK2 cells) and mCCDC11 cell ECVs were internalized under desmopressin regulation. By contrast, uptake of ECVs derived from the kidney juxtaglomerular cell (RG1) was not desmopressin sensitive, despite this being a mouse cell line. This observation was confirmed by flow cytometry of the cells and NTA of the supernatant. The ECVs were of a similar size profile, suggesting that the selectivity of vasopressin-regulated uptake is based on surface molecules within the ECV, such as proteins. Physiologically, the collecting duct would be exposed to proximal tubular ECVs in greater number than ECVs derived from juxtaglomerular cells, and selective uptake may augment proximally to distal tubular signaling. Therapeutically, such selectivity could be a major advantage for harnessing ECVs as a drug delivery system, permitting precision targeting of therapy to treat segment-specific renal tubular disorders.
The cargo of ECVs includes proteins, mitochondrial DNA, and RNA from their cell of origin. Across a range of cell types, it has been reported that ECVs can transfer miRNA into a recipient cell, and this can result in modulation of target mRNA. We used ECVs that contain miR-503 to demonstrate that desmopressin stimulation results in decreased target mRNA expression in mCCDC11 cells. Urinary ECVs contain multiple RNA species, so there is clear potential for signaling both within the kidney and from other organs to kidney tubular cells. However, recent reports have introduced a note of caution; Chevillet et al.24 reported that the content of miRNA in naturally occurring ECVs is very low. Nevertheless, the therapeutic potential of ECVs is high. Our data suggest that ECVs injected IV can freely enter urine and could be engineered to deliver a complex package of RNA and protein that simultaneously targets multiple steps in an intracellular disease pathway.18 These ECVs could be targeted to specific kidney cell types by manipulation of their surface markers and by hormonal activation of target cells.
In conclusion, the uptake of ECVs by kidney collecting duct cells is regulated by vasopressin via intracellular pathways that also mediate the increase in water permeability. Vasopressin regulation occurs in cell lines, primary cells, mice, and possibly humans. This regulated uptake of ECVs can result in intracellular modulation of target mRNA species. Physiologically, ECVs are a fundamental new mechanism of intercellular communication; therapeutically, ECVs represent a novel vehicle by which RNA therapy can be targeted to specific cells for the treatment of kidney disease.
Concise Methods
Cell Culture
The murine cell line (mCCDC11) was a kind gift from Hans-Peter Gaeggeler and Bernard Rossier (University of Lausanne, Lausanne, Switzerland)25 and was grown as per Street et al.12 Briefly, mCCDC11 cells were grown in DMEM–F12 medium, 1:1 (Gibco, Carlsbad, CA), supplemented with 2% FCS (Invitrogen), 1× insulin transferrin selenium solution (Gibco), 100 IU/ml penicillin, and 100 μg/ml streptomycin (Invitrogen), 50 pM dexamethasone (Sigma-Aldrich, St. Louis, MO), 1 nM 3,3,5-triiodo-l-thyronine sodium salt (Sigma-Aldrich) and 10 ng/ml epidermal growth factor (Sigma-Aldrich). Passaging was achieved by two 10-minute washes with 1 mM EDTA in Dulbecco’s modified PBS, followed by incubation in trypsin EDTA solution (Lonza, Basel, Switzerland). The presence of ECVs in FCS would interfere with our study; thus, after confluency, cells were washed twice in Dulbecco’s modified PBS and grown in serum-free media.
For the different cell type experiments, we used ECVs isolated from the supernatant of HK2 and RG1 cell lines. The HK2 cell line was a kind gift from Dr. Kenneth Simpson (University of Edinburgh, United Kingdom). HK2 cells were grown following the same described method as for mCCDC11 cells. The RG1 cell line was grown by supplementing 1:1 DMEM/F12 (Gibco) with 10% heat-inactivated FCS; IFN-γ (Peprotech, London, United Kingdom) at 100 µg/ml; and 1% insulin transferrin selenium containing 1 mg/ml insulin, 0.55 ml/ml human transferrin, and 0.5 µg/ml sodium selenite (Gibco). We added 1× glutamine, 1× penicillin/streptomycin (Life Technologies, Carlsbad, CA) and 1× antioxidants (Sigma-Aldrich) to this, as well as 10 µM Y-27632 (Tocris Bioscience, Bristol, United Kingdom), then filtered them. The cell culture supernatant was removed from either cell type at 70%–80% confluency of the cell layer.
HUVECs and human microvascular endothelial cells (both from Lonza) were grown in endothelial growth medium-2 (endothelial basal medium-2 supplemented with growth factors) and 2% FBS (Lonza). Lipofectamine RNAiMAX (Life Technologies) was used to transfect HUVECs with pre–miR-503 or pre–miR-control (50 nM final concentration) according to the manufacturer's instructions.
Primary Cortical Collecting Duct Isolation
Male rat kidneys were placed in DMEM/Ham’s F12 medium with GlutaMAX (Gibco) containing 10% FBS (Gibco), 100 U/ml penicillin, and 100 μg/ml streptomycin (Gibco). After the perirenal fat was removed, the kidneys were decapsulated and the cortex macroscopically dissected. The cortex was incubated in serum-free isolation medium containing 1 mg/ml collagenase I and IV (Sigma-Aldrich) for 45 minutes at 37°C. The resultant solution was ground and serially sieved to a final filter size of 40 µm. The filtered solution was centrifuged at 27,000g through a 48% Percoll gradient (Sigma-Aldrich), and the highest band of three to four distinct bands was carefully removed. The matter contained within this band was washed and sieved, and the cells were resuspended in DMEM/Ham’s F12 media with GlutaMAX containing the following: 5 μg/ml insulin, 50 nM dexamethasone, 10 ng/ml EGF, 5 μg/ml transferrin, 30 nM sodium selenite, 10 nM triiodothyronine, 100 U/ml penicillin, 100 μg/ml streptomycin, 25 mM sodium bicarbonate, and 10% FBS. The suspension was plated onto collagen-coated 35-mm tissue culture dishes (Corning, Corning, NY) and maintained in an incubator under humidified 5% carbon dioxide atmosphere at 37°C.
Isolation and Fluorescent Loading of ECVs
Culture medium from the cells was vigorously vortexed, then centrifuged at 15,000g for 10 minutes to pellet any cells, large membrane fragments, and other debris. The supernatant was then centrifuged at 200,000g for 60 minutes to pellet ECVs. The pellet was washed with PBS and then recentrifuged at 200,000g for 60 minutes before final suspension in PBS.12
Pelleted ECVs were loaded with Cell Tracker 655 (Invitrogen) following the manufacturer’s protocol. Briefly, pelleted ECVs were incubated with the Cell Tracker 655 conjugate in 200 μl fresh serum-free culture media for 1 hour at 37°C. These labeled ECVs were washed with fresh media before being put back on confluent cells.
Collecting Duct Cell Stimulation
Desmopressin (Sigma-Aldrich) was added as per Street et al.12 For short periods of stimulation, desmopressin and fluorescently loaded ECVs were added together for 1–8 hours. With longer time periods of cell stimulation (24–96 hours), for the final 24 hours of stimulation fluorescent ECVs were added. At the end of the study, the supernatant was collected for NTA analysis and cells were removed by trypsinization (as described previously12) for flow cytometry. In addition to desmopressin, in specific experiments for the final 24 hours the cells were treated with tolvaptan (10 nM) (Sigma-Aldrich), endothelin-1 (10 pM) (Sigma-Aldrich), or H-89 (25 µM) (Sigma-Aldrich).26 Treatment with the Dynamin Inhibitor I (Dynasore 150 μM, Sigma-Aldrich) was for 45 minutes immediately before addition, as per published studies.21 Cells were treated with 10 µM forskolin from Coleus forskohlii (Sigma-Aldrich) and incubated with fluorescent ECVs overnight.27,28 To polarize the mCCDC11 cells, they were cultured on the polyester membrane of Transwell inserts (Corning Costar Co.) at a high density to allow the cells to be confluent within 3 days. Desmopressin was added to the top or bottom Transwell chamber for 48 hours; fluorescent ECVs were subsequently added, then left to incubate for the final 24 hours of the 96-hour period.
Particle Size and Concentration Distribution Measurement with NTA
As per our published method,14 ECVs were analyzed using the NanoSight LM 10 instrument (NanoSight Ltd, Amesbury, United Kingdom). The analysis settings were optimized and kept constant between samples, and each video was analyzed to give the mean, mode, median, and estimated concentration for each particle size. All experiments were carried out at a 1:1000 dilution, yielding particle concentrations in the region of 1×108 particles/ml in accordance with the manufacturer’s recommendations. All samples were analyzed in triplicate. For fluorescent NTA analysis, a 532-nm (green) laser diode excited the fluorescent-loaded ECVs with a long-pass filter (430 nm).
For urine studies, fluorescent labeling with antibody conjugated to Quantum dots was used with NTA. Anti-CD24 antibody was a kind gift of Dr. P. Altevogt (German Cancer Research Center, Heidelberg, Germany). Mouse anti-cubilin and anti-podocalyxin–like protein antibodies were purchased from Abcam, Inc. (Cambridge, MA) and EMD Millipore (Billerica, MA), respectively. Following the manufacturer’s protocol, Qdots were conjugated to antibodies with a Qdot 605 Antibody Conjugation Kit (Invitrogen). For fluorescent NTA analysis, a 532-nm (green) laser diode excited the Qdots with a long-pass filter (430 nm) so that only fluorescent particles were tracked and labeled particle concentration determined by NTA software.
Flow Cytometry for Total Cell Fluorescence
Total cell fluorescence was measured by flow cytometry on a 5LSR Fortessa cytometer (BD Biosciences, San Jose, CA). Cells were briefly stained with 1 μM 4′,6-diamidino-2-phenylindole (DAPI) nucleus stain (Sigma-Aldrich) and having been exposed to Cell Tracker 655 labeled ECVs as described, was excited with a violet laser (405 nm); emission was detected using 450/50 and 630/70 band-pass filters, respectively. Gates were set using unstained cells and cells stained with DAPI alone. Flow cytometry data were analyzed with FlowJo software, version 8 (Ashland, OR), and the results are presented as the percentage of total fluorescent cells.
Fluorescence Microscopy
Using control and desmopressin-stimulated cells grown on a coverslip in a 35-mm glass-bottom dish or fixed mouse tissue, internalization of labeled ECVs with DAPI-stained mCCDC11 cell nuclei and phalloidin-stained cell membranes (Sigma-Aldrich) or AQP2 fluorescent antibodies (Merck Millipore) was visualized by an Olympus AX-70 Provis epifluorescence microscope equipped with a Hamamatsu Orca II CCD camera. Images were collected with a 60× oil immersion objective lens and acquired using MDaemon software (Zenn, Manchester, United Kingdom). Each picture was acquired with laser intensities and amplifier gains adjusted to avoid pixel saturation and were analyzed using Adobe Photoshop CC 2014 (Adobe Systems, Inc., San Jose, CA).
Transmission electron microscopy of mCCDC11 cell culture supernatant was performed as previously described.12
RNA Extraction and Quantitative Real-Time Analysis
Total RNA was extracted using miRNeasy kit (Qiagen, Germantown, MD). Real-time quantification to measure miRNAs was performed with the TaqMan miRNA reverse transcription kit and miRNA assay (miR-503) (Applied Biosystems, Foster City, CA, and Life Technologies) with Lightcycler 480 (Roche Diagnostics, Indianapolis, IN). SYBR quantitative PCR was used to measure vascular endothelial growth factor A, fibroblast growth factor 2, cell division cycle 25A, cyclin-1E, and 18S rRNA. Primers are predesigned from Sigma-Aldrich (KiCqStart Primers). Each reaction was performed in triplicate. Quantification was performed by the 2∆∆Ct method.29
Animals
All experiments were conducted in accordance with the United Kingdom Home Office regulations and the Animals (Scientific Procedures) Act 1986. Wild-type (C57BL6/J and CD1) mice were matched for sex and age across experiments. General anesthesia was induced by intraperitoneal injection of 100 mg/kg thiobutabarbital sodium, and venous access was gained via the jugular vein. Urine flow was maintained throughout with a 0.9% saline infusion (0.2 ml/10 g body wt per hour, IV). A bolus of Cell Tracker–labeled ECVs from mCCDC11 cells was injected in a final volume of 0.1 ml (IV). Urine was collected via a urinary catheter for 2 hours after injection. The injection of ECVs was repeated without or with preceding IV administration of tolvaptan (0.3 mg/kg) or furosemide (1 mg/kg).
Clinical Case Study
Repeated urine samples were obtained from a 16-year-old male patient with stable central diabetes insipidus secondary to a craniopharyngioma, who was being routinely treated with daily desmopressin (dDAVP nasal spray; 0.1 ml [10 μg] desmopressin acetate per spray). The patient samples were initially stored at 4°C, then frozen at −80°C. Analysis of the patient samples were performed by a researcher blinded to the timing of desmopressin treatment. The protocol was approved by the institutional ethical review body, and informed consent was obtained.
Statistical Analyses
Unless otherwise indicated, analyses were performed on data generated from triplicate results. Data were analyzed using GraphPad Prism software, version 6 (GraphPad Software, La Jolla, CA). From the NTA data, the area under the curve was determined following the trapezoidal rule for particles sized 20–100 nm. Flow cytometry data were analyzed using FlowJo, version 8, gating around the main cell population. Total cell fluorescence was determined as the median of single cells lying in a gate defined proportioned from the unstained control, in a forward scatter versus side scatter dot plot, expressed as percentage total cell fluorescence. Nonparametric Kolmogorov–Smirnov t tests and ANOVAs were used to determine significant differences between different conditions. A P value <0.05 was the level of nominal significance.
Disclosures
None.
Experiments were designed and performed by W.O. with assistance from K.M.S., J.R.I., J.M.S., E.E.M., and R.W.H. Supervision was by J.W.D. and M.A.B, supported by P.J.S.L., E.O., A.C., S.J.F., C.D.G., and D.J.W. All authors contributed to the writing of the manuscript.
W.O. was funded by a grant from the Diabetes Research & Wellness Foundation, and K.M.S. and J.R.I. were funded by British Heart Foundation studentships (FS/15/63/32033 and FS/11/78/29328, respectively). J.W.D. acknowledges the support of the National Health Service Research Scotland through National Health Service Lothian. All the authors acknowledge the contribution of the British Heart Foundation Centre of Research Excellence Award (RE/08/001) and the United Kingdom Regenerative Medicine Platform Niche Hub.
References
1. Caby M-P, Lankar D, Vincendeau-Scherrer C, Raposo G, Bonnerot C: Exosomal-like
vesicles are present in human blood plasma. Int Immunol 17: 879–887, 200515908444
2. Lässer C, Alikhani VS, Ekström K, Eldh M, Paredes PT, Bossios A, Sjöstrand M, Gabrielsson S, Lötvall J, Valadi H: Human saliva, plasma and breast milk
exosomes contain RNA: Uptake by macrophages. J Transl Med 9: 9, 201121235781
3. Pisitkun T, Shen R-F, Knepper MA: Identification and proteomic profiling of
exosomes in human urine. Proc Natl Acad Sci USA 101: 13368–13373, 200415326289
4. Vlassov AV, Magdaleno S, Setterquist R, Conrad R:
Exosomes: Current knowledge of their composition, biological functions, and diagnostic and therapeutic potentials.
Biochim Biophys Acta 1820: 940–948, 2012
5. Bramham K, Mistry HD, Poston L, Chappell LC, Thompson AJ: The non-invasive biopsy--will urinary proteomics make the renal tissue biopsy redundant? QJM 102: 523–538, 200919553250
6. Valadi H, Ekström K, Bossios A, Sjöstrand M, Lee JJ, Lötvall JO: Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells. Nat Cell Biol 9: 654–659, 200717486113
7. Tian T, Zhu YL, Hu FH, Wang YY, Huang NP, Xiao ZD: Dynamics of exosome internalization and trafficking. J Cell Physiol 228: 1487–1495, 201323254476
8. Hiemstra TF, Charles PD, Gracia T, Hester SS, Gatto L, Al-Lamki R, Floto RA, Su Y, Skepper JN, Lilley KS, Karet Frankl FE: Human urinary
exosomes as innate immune effectors. J Am Soc Nephrol 25: 2017–2027, 201424700864
9. Alvarez ML, Khosroheidari M, Kanchi Ravi R, DiStefano JK: Comparison of protein, microRNA, and mRNA yields using different methods of urinary exosome isolation for the discovery of kidney disease biomarkers. Kidney Int 82: 1024–1032, 201222785172
10. Borges FT, Melo SA, Özdemir BC, Kato N, Revuelta I, Miller CA, Gattone VH 2nd, LeBleu VS, Kalluri R: TGF-β1-containing
exosomes from injured epithelial cells activate fibroblasts to initiate tissue regenerative responses and fibrosis. J Am Soc Nephrol 24: 385–392, 201323274427
11. Collino F, Bruno S, Incarnato D, Dettori D, Neri F, Provero P, Pomatto M, Oliviero S, Tetta C, Quesenberry PJ, Camussi G: AKI recovery induced by mesenchymal stromal cell-derived extracellular
vesicles carrying microRNAs. J Am Soc Nephrol 26: 2349–2360, 201525901032
12. Street JM, Birkhoff W, Menzies RI, Webb DJ, Bailey MA, Dear JW: Exosomal transmission of functional aquaporin 2 in kidney cortical collecting duct cells. J Physiol 589: 6119–6127, 201122025668
13. Keller S, Rupp C, Stoeck A, Runz S, Fogel M, Lugert S, Hager HD, Abdel-Bakky MS, Gutwein P, Altevogt P: CD24 is a marker of
exosomes secreted into urine and amniotic fluid. Kidney Int 72: 1095–1102, 200717700640
14. Oosthuyzen W, Sime NE, Ivy JR, Turtle EJ, Street JM, Pound J, Bath LE, Webb DJ, Gregory CD, Bailey MA, Dear JW: Quantification of human urinary
exosomes by nanoparticle tracking analysis. J Physiol 591: 5833–5842, 201324060994
15. Knepper MA, Star RA:
Vasopressin: Friend or foe? Nat Med 14: 14–16, 200818180711
16. Kohan DE: The renal medullary endothelin system in control of sodium and water excretion and systemic blood pressure. Curr Opin Nephrol Hypertens 15: 34–40, 200616340664
17. Caporali A, Meloni M, Völlenkle C, Bonci D, Sala-Newby GB, Addis R, Spinetti G, Losa S, Masson R, Baker AH, Agami R, le Sage C, Condorelli G, Madeddu P, Martelli F, Emanueli C: Deregulation of microRNA-503 contributes to diabetes mellitus-induced impairment of endothelial function and reparative angiogenesis after limb ischemia. Circulation 123: 282–291, 201121220732
18. Zhou B, Ma R, Si W, Li S, Xu Y, Tu X, Wang Q: MicroRNA-503 targets FGF2 and VEGFA and inhibits tumor angiogenesis and growth. Cancer Lett 333: 159–169, 201323352645
19. Musch MW, Arvans DL, Wang Y, Nakagawa Y, Solomaha E, Chang EB: Cyclic AMP-mediated endocytosis of intestinal epithelial NHE3 requires binding to synaptotagmin 1. Am J Physiol Gastrointest Liver Physiol 298: G203–G211, 201019926819
20. Macia E, Ehrlich M, Massol R, Boucrot E, Brunner C, Kirchhausen T: Dynasore, a cell-permeable inhibitor of dynamin. Dev Cell 10: 839–850, 200616740485
21. Kirchhausen T, Macia E, Pelish HE: Use of dynasore, the small molecule inhibitor of dynamin, in the regulation of endocytosis. Methods Enzymol 438: 77–93, 200818413242
22. Kohan DE: Biology of endothelin receptors in the collecting duct. Kidney Int 76: 481–486, 200919494799
23. Cheng Y, Wang X, Yang J, Duan X, Yao Y, Shi X, Chen Z, Fan Z, Liu X, Qin S, Tang X, Zhang C: A translational study of urine miRNAs in acute myocardial infarction. J Mol Cell Cardiol 53: 668–676, 201222921780
24. Chevillet JR, Kang Q, Ruf IK, Briggs HA, Vojtech LN, Hughes SM, Cheng HH, Arroyo JD, Meredith EK, Gallichotte EN, Pogosova-Agadjanyan EL, Morrissey C, Stirewalt DL, Hladik F, Yu EY, Higano CS, Tewari M: Quantitative and stoichiometric analysis of the microRNA content of
exosomes. Proc Natl Acad Sci USA 111: 14888–14893, 201425267620
25. Gaeggeler H-P, Gonzalez-Rodriguez E, Jaeger NF, Loffing-Cueni D, Norregaard R, Loffing J, Horisberger JD, Rossier BC: Mineralocorticoid versus glucocorticoid receptor occupancy mediating aldosterone-stimulated sodium transport in a novel renal cell line. J Am Soc Nephrol 16: 878–891, 200515743993
26. Gupta IR, Piscione TD, Grisaru S, Phan T, Macias-Silva M, Zhou X, Whiteside C, Wrana JL, Rosenblum ND: Protein kinase A is a negative regulator of renal branching morphogenesis and modulates inhibitory and stimulatory bone morphogenetic proteins. J Biol Chem 274: 26305–26314, 199910473586
27. Butterworth MB, Edinger RS, Johnson JP, Frizzell RA: Acute ENaC stimulation by cAMP in a kidney cell line is mediated by exocytic insertion from a recycling channel pool. J Gen Physiol 125: 81–101, 200515623897
28. Umenishi F, Narikiyo T, Vandewalle A, Schrier RW: cAMP regulates
vasopressin-induced AQP2 expression via protein kinase A-independent pathway. Biochim Biophys Acta 1758: 1100–1105, 200616844078
29. Schmittgen TD, Livak KJ: Analyzing real-time PCR data by the comparative C(T) method. Nat Protoc 3: 1101–1108, 200818546601
