Functional and molecular evidence for expression of the renin angiotensin system and ADAM17-mediated ACE2 shedding in COS7 cells

Abstract

The renin angiotensin system (RAS) plays a vital role in the regulation of the cardiovascular and renal functions. COS7 is a robust and easily transfectable cell line derived from the kidney of the African green monkey, Cercopithecus aethiops. The aims of this study were to 1) demonstrate the presence of an endogenous and functional RAS in COS7, and 2) investigate the role of a disintegrin and metalloproteinase-17 (ADAM17) in the ectodomain shedding of angiotensin converting enzyme-2 (ACE2). Reverse transcription coupled to gene-specific polymerase chain reaction demonstrated expression of ACE, ACE2, angiotensin II type 1 receptor (AT₁R), and renin at the transcript levels in total RNA cell extracts. Western blot and immunohistochemistry identified ACE (60 kDa), ACE2 (75 kDa), AT₁R (43 kDa), renin (41 kDa), and ADAM17 (130 kDa) in COS7. At the functional level, a sensitive and selective mass spectrometric approach detected endogenous renin, ACE, and ACE2 activities. ANG-(1–7) formation (m/z 899) from the natural substrate ANG II (m/z 1,046) was detected in lysates and media. COS7 cells stably expressing shRNA constructs directed against endogenous ADAM17 showed reduced ACE2 shedding into the media. This is the first study demonstrating endogenous expression of the RAS and ADAM17 in the widely used COS7 cell line and its utility to study ectodomain shedding of ACE2 mediated by ADAM17 in vitro. The transfectable nature of this cell line makes it an attractive cell model for studying the molecular, functional, and pharmacological properties of the renal RAS.

the renin angiotensin system (RAS) plays a crucial role in blood pressure regulation and electrolyte balance (7). The RAS cascade is a multienzymatic system that begins with the formation of angiotensin I (ANG I) from hepatic angiotensinogen via renin (32). Angiotensin converting enzyme (ACE) cleaves the inactive decapeptide ANG I at the COOH-terminal dipeptide (l-histidyl-l-leucine), generating the potent vasopressor octapeptide ANG II (12). ANG II and ACE are involved in the initiation and progression of several diabetic complications such as retinopathy, nephropathy, hypertension, and cardiovascular disease (35). Blockade of the ANG II type 1 receptor (AT₁R) augments renal plasma flow and suppresses filtration fraction in type II diabetic patients suggesting that a RAS blockade improves intrarenal hemodynamics (14). Furthermore, clinical trials have shown that blockade of RAS with ACE inhibitors (ACEIs) and AT₁R blockers (ARBs) decreases the incidence of diabetes mellitus (13).

The recently discovered ACE2, a homologue of ACE, provides a new target site for therapeutic intervention to palliate hypertension and renal diseases (11, 45). ACE2 counterbalances ACE effects by cleaving a single COOH-terminal amino acid of ANG II to form the vasodilator peptide ANG-(1–7), suggesting a renoprotective role for ACE2. Indeed, pharmacological blockade of ACE2 in diabetic mice results in increased albuminuria, glomerular mesangial matrix expansion, and fibronectin deposition (41, 57). Similarly, chronic infusion of ANG-(1–7) has been shown to improve insulin resistance and normalize the hypertensive phenotype as well as cardiac dysfunction and remodeling (17, 34, 39). Previously, significantly decreased renal ACE2 was detected in diabetic animals and in mice with chronic kidney disease (5, 10, 26, 27, 44). However, recent studies provided evidence for increased expression of ACE2 in renal proximal tubules in kidneys of diabetic mice (6, 56, 57) and urine (6, 38, 42, 52), suggesting that urinary ACE2 levels reflect renal pathophysiology. Therefore, ACE2 was proposed as a potential early biomarker of kidney disease. Indeed, increased ACE2 shedding into urine has been described for diabetic mice (6, 38, 52) and for patients with chronic kidney disease, in diabetic renal transplant patients, and in patients with diabetic nephropathy (29, 48, 53).

ACE2 colocalizes in the kidney with a disintegrin and metalloproteinase (ADAM17) (6, 38), a metalloprotease capable of cleaving the ectodomain of ACE2 but not that of ACE (24, 25, 49), suggesting that ADAM17 may be directly involved in the ectodomain shedding of ACE2. Evidence for ACE2 shedding via an ADAM17-mediated pathway has been provided in human proximal tubular HK-2 kidney cells showing that release of ACE2 into the media was blocked by inhibition with a specific ADAM17 inhibitor (38). Moreover, high glucose and ANG II stimulated ACE2 shedding via ADAM17 in a primary culture of mouse proximal tubular cells (54), while silencing of ADAM17 prevented ACE2 shedding in human hepatoma Huh7 cells (31). However, to dissect direct effects of ADAM17 expression and activity on ACE2 shedding and to study the cross talk between the negative and positive regulation of the RAS after pathophysiological stimulus a stable, rapid throughput in vitro assay is needed.

The monkey kidney fibroblast-like COS7 cell line is widely used as a robust line because of its easy propagation and it is amenable to genetic manipulation. The versatility of COS7 cells allows the study of the functional regulation of single genes as well as the functional reconstitution of multiprotein complexes, including components of the RAS (3, 4, 16, 23, 28, 30, 40, 51). Interestingly, COS7 cells functionally express a myriad of genes previously assumed to be absent (1, 2, 21). Therefore, the goal of this study was to examine whether the COS7 cell line endogenously expresses a previously unrecognized, functional RAS and is a suitable model to study ADAM17-mediated ACE2 shedding.

METHODS

Culture of COS7 cells.

COS7 cells (ATCC, Manassas, VA) were grown at 37°C and 5% CO₂ in T75 cm² canted neck, vent cap flasks (Corning, NY) using 10 ml of Dulbecco's modified Eagle's high glucose medium with 10% fetal bovine serum. In some experiments, cells were grown to 90% confluence followed by incubation with serum-free media for 24 h. Cell culture media (10 ml) were collected and centrifuged at 10,000 g for 10 min at 4°C to remove cellular debris and supplemented with 10 μl Complete Lysis-M EDTA-Free buffer (Roche Applied Science, Indianapolis, IN) containing protease inhibitors and 2.5 mmol/l phenylmethanesulphonyl fluoride (PMSF; Sigma-Aldrich, St Louis, MO). To obtain whole cell lysates for Western blot and mass spectrometry analyses, cells were washed with cold PBS and harvested in Complete Lysis-M EDTA-Free buffer containing protease inhibitors and 2.5 mmol/l PMSF. For mass spectrometry experiments, cell culture media were concentrated by ultrafiltration with Amicon Ultra-0.5 10 K Centrifugal Filter Devices (Millipore, Billerica, MA). Total protein content was determined using Bradford reagent (Bio-Rad, Hercules, CA) with bovine serum albumin as a standard.

Reverse transcriptase-polymerase chain reaction.

Total RNA was obtained from early passaged and nearly confluent COS7 cells using the RNeasy Protect Mini kit (Qiagen, Valencia, CA) following the manufacturer instructions. Potential contamination of genomic DNA was eliminated by treating RNA samples with DNAse I (Invitrogen, Carlsbad, CA). First-strand cDNA synthesis and reverse transcription were performed essentially as described in detail elsewhere (8). The sets of primers used to amplify ACE, ACE2, AT₁R, and renin transcripts were designed based on human sequences of reference posted in RefSeq database. Primer sequences were as follows (5′-3′): ACE-593 sense: TCGGCCTGGGACTTCTACAA, ACE-593 antisense: ATGTCACACTTGTGCAGGGG; ACE-528 sense: ACATCCCAGGTGGTGTGGAA, ACE-528 antisense: CAGGGATGGTGTCTCGTACA; ACE2-679 sense: GAG AGA GCA TCT TCA TTG ACA TTG, ACE2-679 antisense: AGC ACT GCT CAA ACA CTG TGA; ACE2-608 sense: CCTAGAACTGAAGTTGAAAAGGCC, ACE2-608 antisense: GTGAGACCAAATACACACTTTCCC; ACE2-599 sense: TGCTGCACAACCTTTTCTGC, ACE2-599 antisense: GCAGTGGCCTTACATTCATGTTCT; ACE2-567 sense: GAACATCTTCATGCCTATGTGAGG, ACE2-567 antisense: GGGTGACAGAAGACCAATGGATTT; AT₁R-575 sense: TTG CCA GCT ATA ATC CAT CG, AT₁R-575 antisense: GGC TTC TTG GTG GAT GAG CT; Renin-566 sense: GCCTGTGTGTATCACAAGCTCT, Renin-566 antisense: GCCTCCATGAGCTTCTCTATG. GAPDH mRNA levels were determined as internal control. The ACE2 and ACE primers were also used to determine the nucleotide sequence of the ACE2 and ACE PCR fragments produced by using reverse transcriptase-polymerase chain reaction (RT-PCR). Sequence analysis of the ACE2 and ACE RT-PCR fragments was performed in at least three separate RT-PCR amplicons obtained from different RNA samples. Sequences of fragments were aligned and matched to human sequences of reference (RefSeq) in silico using Geneious software suite R7 (Biomatters, Auckland, NZ; www.geneious.com).

Immunofluorescence.

COS7 cells were grown to 70% confluence on glass coverslips placed in six-well plates, fixed in absolute methanol, and permeabilized with 4% paraformaldehyde containing 0.25% TritonX 100. After 20 min of incubation at 4°C, cells were washed with cold PBS solution and blocked using 3% normal donkey serum for 1 h at 4°C. This was followed by incubation with primary antibodies: goat anti-ACE (1:250; cat. no. sc-12187; Santa Cruz Biotechnology, Santa Cruz, CA), goat anti-ACE2 (1:250; cat. no. sc-21834; Santa Cruz Biotechnology), goat anti-renin (1:250; cat. no. sc-27318; Santa Cruz Biotechnology), rabbit anti-AT₁R (1:400; cat. no. sc-1173; Santa Cruz Biotechnology), rabbit anti-ADAM17 (1:200; cat. no. ADI-905-249-100; Enzo Life Sciences, Farmingdale, NY), and rabbit anti-mannosidase II (1:100; cat. no. LS-B9328, LifeSpan Biosciences, Seattle, WA) diluted in 3% normal donkey serum overnight at 4°C. Cells were subsequently washed with cold PBS and incubated with secondary CY3-conjugated donkey anti-goat, CY3-conjugated donkey anti-rabbit, or Alexa Fluor 488-conjugated donkey anti-rabbit (Jackson Immunoresearch) antibodies diluted 1:500 with PBS for 2 h at 4°C. Cells were washed, air dried, and coverslip mounted over a drop of Vectashield (Vector Laboratories, Burlingame, CA). To test the specificity of the rabbit anti-AT₁R antibody, the primary antibody was combined with five times its concentration of blocking peptide (sc-1173 P; Santa Cruz Biotechnology) in 500 μl of PBS. The peptide sequence used for making the antibody epitope lies within the NH₂-terminal extracellular domain (amino acid 1–50) of human AT₁R. To ensure functionality of the preincubated antibody, AT₁R antibody was also incubated alone in 500 μl of PBS without the addition of blocking peptide. The AT₁R antibody/peptide-PBS and AT₁R antibody-PBS mixtures were incubated overnight at 4°C followed by dilution with normal donkey serum to a final concentration of 3% and incubation of cells overnight at 4°C. Cells were washed with cold PBS and incubated with 1:500 diluted secondary CY3-conjugated donkey anti-rabbit IgG for 2 h at 4°C. Cells were washed, air dried, and coverslip mounted over a drop of Vectashield. Images were taken with SPOT scope microscope.

Western blot.

Protein expression of renin, ACE, ACE2, and AT₁R proteins in mouse kidney homogenates, COS7 cell lysates, and media was determined using 10% SDS-PAGE. Equal volumes of cell lysates or media and loading buffer (8% SDS, 125 mmol/l Tris·HCl, pH 6.8, 20% glycerol, 0.02% bromophenol blue, and 100 mmol/l dithiothreitol) were mixed and boiled for 6 min. After boiling, protein equivalent samples (2 μg protein for COS7 media, 30 μg protein for COS7 lysates, and 30 μg for kidney homogenates) were loaded and separated by using a 10% SDS-PAGE system. Separated proteins were transferred onto a 0.2-μm polyvinylidene difluoride membrane (Millipore) using a Bio-Rad transfer apparatus for 2 h. The membrane was subsequently blocked with 10% nonfat powdered milk in TBST for 1 h at room temperature and probed with diluted rabbit anti-ACE2 (1:250; cat. no. sc-20998; Santa Cruz Biotechnology), goat anti-ACE (1:250; Santa Cruz Biotechnology, CA), goat anti-renin (1:250; Santa Cruz Biotechnology), rabbit anti-AT₁R (1:500; Santa Cruz Biotechnology), rabbit anti-ADAM17 (1:1,000; Cell Signaling), or mouse anti-β-actin (1:4,000; Sigma) primary antibodies for 2 days at 4°C. The membranes were washed and incubated with horseradish peroxidase-conjugated donkey anti-rabbit (1:20,000; Jackson Immunoresearch), donkey anti-goat (1:2,000; R&D, Minneapolis, MN), donkey anti-mouse (1:40,000; Jackson Immunoresearch), or goat anti-rabbit (1:2,000; Santa Cruz Biotechnology) secondary antibodies at room temperature for 1 h. To test the specificity of the rabbit anti-AT₁R antibody, the primary antibody was combined with five times its concentration of blocking peptide sc-1173 P in 500 μl of PBS. To ensure functionality of the preincubated antibody, AT₁R antibody was also incubated alone in 500 μl of PBS without addition of blocking peptide. The AT₁R antibody/peptide-PBS and AT₁R antibody-PBS mixtures were incubated overnight at 4°C followed by dilution with 5% nonfat powdered milk in TBST and incubation of Western blots for 2 days at 4°C. The membranes were washed and incubated with horseradish peroxidase-conjugated secondary antibodies diluted 1:20,000 in TBST at room temperature for 1 h. Blots were visualized using supersignal chemiluminescent substrate and a Fujifilm image analyzer (LAS 3000; Image quant).

Mass spectrometric detection of enzymatic activity in lysates and cell culture media.

Enzymatic RAS activities were measured using matrix-assisted laser desorption/ionization (MALDI) mass spectrometry as described with some modifications (12, 18, 19). For renin activity, COS7 cell lysates (∼10–20 μg protein) were incubated for 5 min at 37°C in 25 mM bicine buffer pH 7.6 containing Complete Lysis-M EDTA-Free protease inhibitors, 2.5 mM PMSF, and 1 μg tetradecapeptide. For renin inhibitor studies, the assay was incubated with 100 μM renin inhibitor [Z-Arg-Arg-Pro-Phe-His-Sta-Ile-His-Lys-(Boc)-OMe] obtained from Bachem Bioscience (King of Prussia, PA). For ACE activity, COS7 cell lysates (∼10–20 μg protein) were incubated for 60 min at 37°C in 50 mM MES buffer pH 6.75 containing 2 mM PMSF and 100 ng ANG I. For ACE inhibitor studies, the assay was incubated with 100 μM captopril. For ACE2 activity, COS7 cell lysates (∼10–20 μg protein) were incubated for 60 min at 37°C in 25 mM bicine buffer pH 7.6 containing Complete Lysis-M EDTA-Free protease inhibitors, 2.5 mM PMSF, 100 ng ANG II, and 500 μM 4-aminophosphonobutyric acid (4-APBA) for aminopeptidase A inhibition. Concentrated cell media (∼50 μg protein) were incubated for 120 min at 37°C in 25 mM bicine buffer pH 7.6 containing Complete Lysis-M EDTA-Free protease inhibitors, 2.5 mM PMSF, and 100 ng ANG II. For ACE2 inhibitor studies, the assay was spiked with 100 μM MLN-4760 (MLN) for ACE2 inhibition or 100 μM Z-pro-prolinal (ZPP) for prolyl carboxypeptidase/prolyl endopeptidase inhibition. The reaction was stopped by acidification with trifluoroacetic acid (TFA; final concentration 1%). The samples were diluted 1:10 in 90% acetonitrile containing 0.3% TFA. Mass spectra were obtained using an Autoflex III smartbeam MALDI TOF/TOF instrument (Bruker Daltonics). The mass spectrometer was operated with positive polarity in reflectron mode. A total of 10,000 laser shots was acquired randomly for each spot in the range of m/z 500-3,000 at a laser frequency of 100 Hz. Spectra were mass calibrated by collecting 200 laser shots of spots containing Bruker peptide calibration standard II consisting of nine peptide standards covering a mass range of 700-3,200 Da. Signals for peptide products were fragmented using the Bruker Lift method and identified upon comparison to standard peptides.

Stable transfection and selection.

The procedures followed to produce clones of COS7 cells stably expressing short hairpin (sh)RNAs were essentially as described previously (9) with slight modifications. Briefly, confluent COS7 cells routinely grown in T75 plates were harvested by trypsin digestion and seeded on six-well plates to attain an approximate confluence of 75%. After 16 h, cells were washed three times with PBS and transfected with 2.5 μg of pGIPz-GFP.shADAM17 (Open Biosystems, Lafayette, CO) using Lipofectamine2000 (Invitrogen) according to the manufacturer's instructions. Two days posttransfection, cells were washed and observed under an inverted fluorescence microscope to identify GFP-expressing cells and to estimate transfection efficiency, which in our hands was ∼30%. Then, fully supplemented media containing 2.5 μg/ml puromycin (InvivoGen, San Diego, CA) were added to start the selection of individual clones. The selection process consisted of replacing puromycin media each day for at least 3 wk. Once confluence was reached, cells in each well were harvested, diluted 1:1,000, and replated in 24-well plates until confluence. Wells with most of the GFP-expressing cells were once more collected, diluted, and reseeded onto 24-well plates until clear single colonies could be observed under the inverted fluorescence microscope. Individual GFP-positive and puromycin-resistant colonies were scraped-off plates and seeded in 25-cm² flasks to identify isolated colonies of cells. These colonies were gently scrapped and propagated in 75-cm² flasks. Upon confluence, each flask was split into four flasks for permanent stocks and further characterization.

ADAM17 enzymatic activity.

ADAM17 activity was measured using an internally quenched fluorogenic substrate Mca-PLAQAV-Dpa-RSSSR-NH2 (Enzo Life Sciences) according to the manufacturer's instructions with some modifications. Samples (20–30 μg) were incubated with the assay buffer (50 mM Tris, 5 mM ZnCl₂, 150 mM NaCl, and 10 mM lisinopril) with pH 9 and 50 μM Mca-PLAQAV-Dpa-RSSSR-NH2 for 5–6 min. Fluorescence was measured at an excitation of 328 nm and emission of 393 nm using a Fusion Packard plate reader.

Statistical analysis.

Statistical analysis was performed using GraphPad Prism 5.01. All data are expressed as means ± SE. Unpaired Student's t-test and Mann-Whitney post-test were used to evaluate ADAM17 and ACE2 in COS7 cell lysates. A value of P < 0.05 was considered statistically significant.

RESULTS

Expression of RAS at the transcript level in COS7 cells.

To ascertain molecular identity and to demonstrate endogenous RAS expression in COS7 cells at the transcript level we used RT-PCR coupled to human-specific primer sets designed to amplify mRNA's encoding for RAS components. Figure 1 suggests that the transcripts for AT₁R, renin, ACE2, and ACE are present in COS7 cells. To demonstrate identity, the RT-PCR products obtained for ACE2 and ACE were sequenced in both directions. As shown in Fig. 1, the nucleotide sequences corresponding to ACE2 and ACE are >95% identical to human ACE2 and ACE cDNAs. Together, these results suggest that the RAS system is expressed at the transcript level in COS7 cells.

Fig. 1.

Expression of renin angiotensin system (RAS) mRNAs in COS7 cells. A: transcripts were detected by RT-PCR using COS7 total RNA as template and gene-specific primers for PCR-amplification. As negative control, water was used instead of RNA. Total RNA was reverse transcribed to cDNA and angiotensin II type 1 receptor (AT₁R), renin, angiotensin converting enzyme-2 (ACE2), ACE, and GAPDH mRNAs were detected by PCR. Each primer set was designed to produce amplicons of 566 bp for renin; 575 bp for AT₁R; 679, 608, 599, and 567 bp for ACE2; and 593 and 528 bp for ACE, respectively. The negative control was prepared using water and the GAPDH primer set. The red asterisk indicates the amplicon that was used to determine the nucleotide sequence of the ACE2 and ACE PCR fragments produced in RT-PCR reactions. B: molecular identity of ACE2 amplicon (actual sequence chromatogram) is shown together with the predicted genomic organization of the human ACE2 gene (NM_021804). C: molecular identity of ACE amplicon (actual sequence chromatogram) is shown together with the predicted transcript organization of human ACE mRNA (NM_000789.2).

Protein expression of RAS components in COS7 cells.

To corroborate the previous findings at the protein level, the immunological presence of the RAS components in COS7 cell lysates was ascertained by using Western blotting. Figure 2 demonstrates a qualitative analysis of RAS components renin (41 kDa), ACE (60 kDa), ACE2 (60 kDa), and AT₁R (43 kDa) expressed in COS7 cells. Since AT₁R antibodies are known to be unspecific (20), we incubated Western blots with AT₁R antibody in presence of the blocking peptide directed against the AT₁R antibody to show the specificity of the AT₁R antibody used in the Western blot studies. AT₁R was not detected in Western blots incubated with AT₁R blocking peptide. The loading control β-actin was detected at 42 kDa. While the size of COS7 AT₁R and renin was comparable to the size observed in mouse kidney homogenates, a 195-kDa band was not detected for ACE and a 100-kDa band was not detected for ACE2 in COS7 lysates compared with mouse kidney homogenates.

Fig. 2.

Western blot analysis of RAS components in mouse kidney and COS7 cell lysates. Proteins were separated on a 10% SDS-PAGE and immunoblotted using renin, ACE, ACE2, AT₁R, and β-actin antibodies. Immunoreactive bands specific for renin (41 kDa), ACE (60 kDa), ACE2 (75 kDa), AT₁R (43 kDa), and β-actin (42 kDa) were detected in COS7 cells. KO, knockout.

Immunofluorescence detection of RAS components in COS7 cells.

To further substantiate the concept that RAS components are present in COS7 cells, immunohistochemistry was performed. As shown in Fig. 3A, protein expression of renin, ACE, and AT₁R was confirmed in COS7 cells. To show specificity of the used AT₁R antibody in the immunofluorescence studies, we added control experiments using the antibody-specific AT₁R blocking peptide (Fig. 3B). Renin was localized in the cytoplasm and ACE in the nucleus. Nuclear and cell surface staining was observed for AT₁R. ACE2 was localized in the perinuclear region of COS7 cells and colocalized with mannosidase II, a marker of the Golgi apparatus (Fig. 4).

Fig. 3.

Detection of renin, ACE, and AT₁R in COS7 cells using immunofluorescence microscopy. A: renin was localized in the cytoplasm, ACE was localized in the nucleus, and AT₁R was localized in the nucleus and cell surface. B: specificity of the AT₁R antibody was tested in incubations with the blocking peptide directed against the AT₁R antibody. The peptide sequence used for making the antibody epitope lies within the NH₂-terminal extracellular domain (amino acid 1–50) of human AT₁R.

Fig. 4.

Detection of ACE2 in COS7 cells using immunofluorescence microscopy and colocalization of ACE2 with Golgi marker mannosidase II.

Renin, ACE, and ACE2 activities in COS7.

A sensitive and selective mass spectrometric approach was used to determine endogenous RAS enzymatic activity in COS7 cell lysates and media. This technique allows for the detection of ANG products formed in cells and media incubated with the natural RAS peptide substrates. Renin activity was detected by incubating COS7 lysates with the renin substrate tetradecapeptide (TDP) (m/z 1,758) and showing the formation of ANG I (m/z 1,296; Fig. 5A). Renin activity was blocked by the addition of renin inhibitor (Fig. 5A). ACE activity was detected by showing formation of ANG II (m/z 1,046) in incubations of COS7 lysates with ANG I (m/z 1,296; Fig. 5B). Addition of the ACE inhibitor captopril reduced ANG II formation in COS7 lysates (Fig. 5B). Shown in Fig. 5C is the formation of ANG-(1–7) in COS7 lysates incubated with ANG II. ANG-(1–7) formation in lysates was not inhibited by the ACE2 inhibitor MLN, while ZPP, a prolyl carboxypeptidase/prolyl endopeptidase inhibitor, markedly reduced ANG-(1–7) formation. Dual blockade using MLN and ZPP completely abolished ANG-(1–7) formation in cell lysates. In contrast, ANG-(1–7) was formed in media from ANG II and was completely blocked by the ACE2 inhibitor MLN but not ZPP (Fig. 5D). Dual blockade with MLN and ZPP completely abolished ANG-(1–7) formation in media most likely due to full inhibition with MLN. The molecular identity of enzymatically produced ANG-(1–7) was confirmed by tandem mass spectrometry (MS/MS) upon comparison with standard compound (Fig. 5E). Taken together, these results demonstrate that the RAS is actively expressed in COS7 cells and ACE2 is shed into the media.

Fig. 5.

Mass spectrometric analysis of RAS components in COS7. A: ANG I formation from tetradecapeptide (TDP) in COS7 lysates in absence or in presence of 100 μM renin inhibitor (RI). B: ANG II formation from ANG I in COS7 lysates in absence or in presence of 100 μM ACE inhibitor captopril. C: ANG-(1–7) formation from ANG II in COS7 lysates in absence or in presence of ACE2 inhibitor MLN-4760 (MLN; 100 μM) or prolyl carboxypeptidase/prolyl endopeptidase inhibitor Z-pro-prolinal (ZPP; 100 μM) or a combination of both. D: ANG-(1–7) formation from ANG II in COS7 media in absence or in presence of ACE2 inhibitor MLN (100 μM) or prolyl carboxypeptidase/prolyl endopeptidase inhibitor ZPP (100 μM) or a combination of both. E: tandem mass spectrometry (MS/MS) of ANG-(1–7) enzymatically produced by COS7 and ANG-(1–7) standard compound.

shRNA-mediated stable silencing of ADAM17.

To demonstrate a direct involvement of ADAM17 in the regulation of ACE2 as suggested by previous studies in our laboratory (6, 38), ADAM17 expression was stably silenced in COS7 cells by using multicystronic lentiviral constructs harboring GFP expression cassettes, selection markers, and shRNAs directed against human ADAM17. Analysis of GFP expression in selected COS7 cells stably transfected with these constructs showed a high percentage of successfully transfected cells (Fig. 6A). ADAM17 expression in stably transfected cells was analyzed by Western blot and showed a nearly complete silencing of ADAM17 expression in these cells upon quantitative analysis of the ratio of ADAM17 to β-actin (Fig. 6B) a finding that correlated with decreased ADAM17 enzymatic activity (Fig. 6C). These results were further corroborated in the immunohistochemical analysis showing a complete downregulation of ADAM17 in silenced cells (Fig. 7). Together, these results demonstrate that not only ADAM17 is expressed in COS7 cells but also that its expression and function can be efficiently silenced.

Fig. 6.

Short-hairpin (sh)RNA-mediated stable silencing of a disintegrin and metalloproteinase-17 (ADAM17). A: fluorescence analysis of GFP expression in COS7 cells transfected with pGIPz-shRNAs against ADAM17. B: Western blot analysis and quantitation of ADAM17 expression in normal and silenced COS7 cells (*P < 0.0001). C: ADAM17 enzymatic activity in normal and silenced COS7 cells (#P < 0.05).

Fig. 7.

Immunofluorescence microscopy of ADAM17 expression in normal and silenced cells.

Ectodomain shedding of ACE2 from COS7 cells.

To gain insights into the mechanisms of ACE2 shedding mediated by ADAM17, ACE2 was analyzed in COS7 lysates and media of normal cells and ADAM17 silenced cells. There was no difference of ACE2 protein expression normalized to β-actin in lysates of silenced cells compared with normal cell lysates (Fig. 8A). In media, ACE2 protein expression was significantly reduced in sh.ADAM17 compared with media of normal cells (Fig. 8B). These results suggest that ADAM17 is directly involved in the ectodomain shedding of ACE2.

Fig. 8.

Western blot analysis of ACE2 in COS7 lysates and cell culture media of normal and silenced cells. A: expression of ACE2 in lysates of normal and silenced cells. B: expression of ACE2 in media of normal and silenced cells (*P < 0.05).

DISCUSSION

Previous studies have suggested that COS7 cells lack endogenous AT₁R, AT₂R, and Mas receptors (3, 4, 16, 23, 28, 30, 40, 51). However, basal expression and activity of MAS and AT₁R were detected in nontransfected or vector-control transfected COS7 cells (16, 40, 51). Therefore, in the present study, we tested the hypothesis that COS7 cells express a previously unrecognized, functional RAS. Indeed, Western blot and immunofluorescence microscopy demonstrated that COS7 cells endogenously express RAS components including ACE, ACE2, AT₁R, and renin making it an attractive in vitro model for studying the molecular, functional, and pharmacological properties of the renal RAS.

The immunoreactive bands for the RAS components renin and AT₁R detected in mouse kidneys were identical to the bands observed in COS7 cells. The bands for ACE2 and ACE at lower molecular mass were comparable between mouse kidney and COS7 cells; however, bands for ACE2 and ACE at higher molecular mass were absent in COS7. While mouse kidney and COS7 cells showed a 60-kDa immunoreactive band for ACE, the 195-kDa band detected in kidney homogenates for ACE was missing in COS7 cells. This discrepancy may be due to distinct distribution patterns of different ACE isoforms identified to-date: a somatic form at high molecular mass, a COOH-terminal isoform at around 90–100 kDa, and a soluble isoform derived from the membrane-bound form (37). In addition, shorter isoforms of both ACE and ACE2 have been associated with increased shedding in diabetic patients and mice. Therefore, it is likely that the shorter fragments of ACE and ACE2 observed in COS7 cells may be due to feedback mechanisms to the growth conditions in high glucose medium.

Evidence for an endogenous, functional RAS was provided using mass spectrometry. Incubation of COS7 cells with ANG I (m/z 1,296) or TDP (m/z 1,758) resulted in formation of ANG II (m/z 1,046) or ANG I (m/z 1,296), respectively. Our results also show that ANG-1-7 (m/z 899) was formed from ANG II (m/z 1,046) in cell lysates and media. Mass spectrometry has previously been used to demonstrate functional RAS in media of cultured mouse podocytes (46) and human glomerular endothelial cells (47). ANG II formation from ANG I was blocked by captopril, suggesting the presence of functional ACE in COS7 cells. Renin activity was confirmed by complete inhibition of ANG I formation from TDP using renin inhibitor. For COS7 media, the ANG-(1–7) formation was solely due to ACE2 as ANG-(1–7) generation was completely blocked in the presence of the selective ACE2 inhibitor MLN. While addition of ZPP, a prolyl carboxypeptidase/prolyl endopeptidase inhibitor, had no effect on ANG-(1–7) formation in media, ZPP markedly reduced ANG-(1–7) formation in COS7 lysates. This finding suggests that alternative, ACE2-independent pathways are involved in the processing of ANG II and that in COS7 cell lysates ACE2 may not be as important in the cleavage of ANG II as it is in media under the tested conditions. In addition to ACE2, ANG-(1–7) can be formed by prolyl endopeptidase (50, 58), prolyl carboxypeptidase (43), neprilysin (50, 55), thimet oligopeptidase (36), and neurolysin (36). The present study indicates a high abundance of alternative ANG-(1–7)-forming enzymes, most likely prolyl carboxypeptidase or prolyl endopeptidase, in COS7 cell lysates but not cell media. Indeed, our laboratory recently identified compensatory ANG-(1–7) formation by prolyl carboxypeptidase in mouse kidneys (19). Involvement of prolyl carboxypeptidase or prolyl endopeptidase in renal ANG-(1–7) formation was also shown in human glomerular endothelial cells (47). Since the ACE2 reaction mixtures were measured at a pH of 7.6, it is very likely that ANG-(1–7) formation in COS7 lysates is mainly due to action of prolyl endopeptidase and not prolyl carboxypeptidase. The lower ACE2 activity in cell lysates compared with cell media could be a result of the COS7 culture conditions in high glucose medium. Previous studies showed increased ACE2 shedding in diabetic mice (6, 38, 52), in diabetic renal transplant patients, in patients with diabetic nephropathy (29, 48, 53), and in primary cultures of mouse proximal tubular cells upon stimulation with 25 mM d-glucose compared with 7.8 mM d-glucose or l-glucose (54).

Ectodomain shedding of soluble ACE2 has been described in several studies including human proximal tubular HK-2 cells (38), mouse proximal tubular primary cells (54), human hepatoma cell lines Huh1 and Huh7 (15), and ACE2-transfected Chinese hamster ovary cells (CHO) (22). ADAM17, a sheddase of tumor necrosis factor-α and several integral proteins, is expressed in the proximal tubules of the mouse kidney (6, 38) and was found in numerous cell cultures including CHO, human proximal tubular HK-2 cells, and Huh (31, 33, 38). ADAM17 is activated by hyperglycemia in diabetic animals (15) and upon stimulation with high glucose (54), which leads to the upregulation of NOX and reactive oxygen species (15). Our previous studies show that the expression of ADAM17 has been confirmed in diabetic nephropathy in both Akita and db/db mice showing colocalization with ACE2 in renal proximal tubules (6, 38). By exploitation of the easily transfectable nature of this cell line, evidence for a direct involvement of ADAM17 in the shedding of ACE2 was provided in this study using shRNA-mediated silencing. Silencing of ADAM17 resulted in a significant decrease of ADAM17 protein expression and activity in COS7 cells. While ADAM17 expression was silenced by 90%, residual ADAM17 protein was sufficient to cause an only modest, although significant drop of ADAM17 enzymatic activity by 40%. Accordingly, ADAM17-silenced cells showed a comparable 40% reduction of ACE2 shedding into cell culture media. New evidence suggests that proteolytic cleavage of ACE2 by ADAM17 is induced by ANG II and mediated through AT₁R (31) suggesting a protective role of AT₁R blockers and ACE inhibitors in kidney disease.

To the best of our knowledge, this study is the first to demonstrate that COS7 cells endogenously express RAS components including ACE, ACE2, AT₁R, and renin. The results challenge previous studies suggesting the lack of RAS expression in COS7 and question the need for transfecting COS7 to study the RAS. The amenable nature of the COS7 cell line makes it an attractive in vitro model for studying ectodomain shedding of ACE2 mediated by ADAM17 along with the molecular, functional, and pharmacological properties of the renal RAS and potential interaction and cross talk between the RAS and other systems.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Fellowship F32-DK-093226 (to N. Grobe), American Diabetes Association Grant JF112 (to M. Di Fulvio), Boonshoft School of Medicine Seed Grant Program (to M. Di Fulvio), American Heart Association Grant SDG 0735112N (to K. M. Elased), and Boonshoft School of Medicine Emerging Science Seed Grant (to K. M. Elased).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

N.G., M.DF., and K.M.E. conception and design of research; N.G., M.DF., N.K., H.C., H.K.S., and R.S. performed experiments; N.G., M.DF., N.K., H.C., H.K.S., R.S., and K.M.E. analyzed data; N.G., M.DF., N.K., H.C., H.K.S., R.S., and K.M.E. interpreted results of experiments; N.G., M.DF., N.K., H.C., H.K.S., and K.M.E. prepared figures; N.G., M.DF., N.K., and K.M.E. drafted manuscript; N.G., M.DF., N.K., and K.M.E. edited and revised manuscript; N.G., M.DF., N.K., H.C., H.K.S., R.S., and K.M.E. approved final version of manuscript.