Chymase-Dependent Generation of Angiotensin II from Angiotensin-(1-12) in Human Atrial Tissue

Since angiotensin-(1-12) [Ang-(1-12)] is a non-renin dependent alternate precursor for the generation of cardiac Ang peptides in rat tissue, we investigated the metabolism of Ang-(1-12) by plasma membranes (PM) isolated from human atrial appendage tissue from nine patients undergoing cardiac surgery for primary control of atrial fibrillation (MAZE surgical procedure). PM was incubated with highly purified 125I-Ang-(1-12) at 37°C for 1 h with or without renin-angiotensin system (RAS) inhibitors [lisinopril for angiotensin converting enzyme (ACE), SCH39370 for neprilysin (NEP), MLN-4760 for ACE2 and chymostatin for chymase; 50 µM each]. 125I-Ang peptide fractions were identified by HPLC coupled to an inline γ-detector. In the absence of all RAS inhibitor, 125I-Ang-(1-12) was converted into Ang I (2±2%), Ang II (69±21%), Ang-(1-7) (5±2%), and Ang-(1-4) (2±1%). In the absence of all RAS inhibitor, only 22±10% of 125I-Ang-(1-12) was unmetabolized, whereas, in the presence of the all RAS inhibitors, 98±7% of 125I-Ang-(1-12) remained intact. The relative contribution of selective inhibition of ACE and chymase enzyme showed that 125I-Ang-(1-12) was primarily converted into Ang II (65±18%) by chymase while its hydrolysis into Ang II by ACE was significantly lower or undetectable. The activity of individual enzyme was calculated based on the amount of Ang II formation. These results showed very high chymase-mediated Ang II formation (28±3.1 fmol×min−1×mg−1, n = 9) from 125I-Ang-(1-12) and very low or undetectable Ang II formation by ACE (1.1±0.2 fmol×min−1×mg−1). Paralleling these findings, these tissues showed significant content of chymase protein that by immunocytochemistry were primarily localized in atrial cardiac myocytes. In conclusion, we demonstrate for the first time in human cardiac tissue a dominant role of cardiac chymase in the formation of Ang II from Ang-(1-12).

Striking differences in staining of cardiomyocytes between Ang-(1-12) peptide block and unblocked tissue slides were noted ( Figure 1). When viewed at higher magnification, almost all atrial cardiac myocytes showed some degree of positive staining with the Ang-(1-12) antibody ( Figure 1).
The chromatograms depicting the effect of peptidase inhibition on 125 I-Ang-(1-12) metabolism are shown in Figure 2-panels A to D. In the presence of all RAS inhibitors, the chromatogram shows a large peak area of unmetabolized Ang-(1-12) along with very small peaks of Ang I and Ang II (Figure 2-panel A), whereas in the absence of all RAS inhibitors the peak area of unmetabolized Ang-(1-12) is a small fraction of that observed in the presence of all inhibitors. The major peak corresponds to Ang II while peaks corresponding to the synthetic forms Ang I, Ang-(1-7), and Ang- (1)(2)(3)(4) are expressed minimally in the absence of all RAS inhibitors ( Figure 2-panel B). Ang peptides formation from 125 I-Ang-(1-12) is only 261% (Ang II) when only lisinopril is removed from the cocktail of inhibitors ( Figure 2-panel C), whereas the selective removal of chymostatin from the cocktail of inhibitors has a diametrically opposite effect. As shown in Figure 2-panel D, Ang II formation from 125 I-Ang-(1-12) metabolism represents now the major peak accounting for 65618%. Removal of chymostatin is also associated with the presence of a minor peak for Ang I (462%). The proportion of Ang II produced from 125 I-Ang- (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12) in the absence of chymostatin (Ang II formation 65618%; P,0.05; n = 9) was not significantly different than that observed when 125 I-Ang-(1-12) was incubated with human atrial membranes in the absence of all inhibitors (Ang II formation 69621%; n = 9) ( Table 1).

Chymase Expression in Human Atrial Tissue
Chymase protein expression, determined in 8 of the 9 human samples using a primary monoclonal anti-human chymase antibody, showed an immunoreactive band (,30 KD) for chymase protein ( Figure 4). The ir-chymase bands were normalized with bactin and relative chymase protein expression in the human atrial tissue was consistence with the HPLC results of chymase activity to generate Ang II products from 125 I-Ang-(1-12) ( Figure 3). In addition, a significant correlation was found between chymase activity and protein expression (r = 0.954, P,0.002).
Immunolocalization of chymase was also analyzed in these atrial tissues by using an anti-mast cell chymase monoclonal antibody. As illustrated in Figure 5, a strong immunostaing for the presence of chymase was detected within the cardiac myocytes of atrial tissue. Negative controls without primary antibody show no staining for chymase ( Figure 5).
A critical role for chymase in accounting for the processing of Ang-(1-12) is reinforced through the corresponding consistency of the chymase activity measured in each of the human samples. Since the atrial tissue employed in these studies was obtained from subjects undergoing cardiac surgery for the treatment of heart rhythm disorders (Table 2), we cannot exclude the possibility that the relatively high chymase activity reflects increased expression of this enzyme in this condition. Although an increase in cardiac chymase activity has been reported in both clinical and experimental forms of heart failure [19][20][21][22][23], several studies suggest that in humans Ang II formation from Ang I is primarily dependent upon chymase rather than ACE [19,20,22,[24][25][26][27]. A comparative study by Akasu et al. [28] in human, hamster, rat, rabbit, dog, pig, and marmoset tissues demonstrated a primary role of lung ACE in all species but human cardiac and lung tissues. In these tissues, a chymase-like activity solely accounted for the Ang II-forming activity. Balcells et al. [29] reported that chymase rather than ACE was the predominant enzyme accounting for intracardiac Ang II generation in humans, whereas Ang II forming activity from ACE was higher in cardiac tissue obtained from mouse, dog, rabbits, rat, and mice. The rodent heart might be an exemption since Wei et al. [27] showed that mast-cell derived chymase contributes to Ang II from Ang I. In their studies the content of Ang II in left ventricular interstitial fluid was not altered by high doses of an ACE inhibitor but was suppressed in the presence of a chymase inhibitor [27]. It is possible that upregulation of chymase expression or activity may occur in conditions of ACE inhibition, since chronic ACE inhibition causes a marked bradykinin/B2 receptor-mediated increase in cardiac interstitial fluid chymase activity that does not occur in mast celldeficient KitW/KitW-v mice [27]. Becari et al. [30] reported an 80% inhibition of Ang I induced carotid artery vasoconstriction by the administration of chymostatin in enalapril-treated SHR rats. Since the atrial tissues were obtained from 7 of the 9 subjects not treated with ACE inhibitors and none were treated with Ang II receptor blockers (Table 2), our findings imply a major and independent role of cardiac chymase in the biotransformation of 125 I-Ang-(1-12) to angiotensin peptides. Our results and those discussed above indicate that species differences may influence the contribution of ACE and chymase to cardiac Ang II formation. Urata et al. [24][25][26]31] first showed in human cardiac ventricles that formation of Ang II by a serine proteinase-dependent enzyme accounted for 80% of Ang II formation while less than 10% of the Ang II produced depended upon ACE. Later studies showed that this serine proteinase was structurally characterized as a human member of the chymase group of enzyme which is not inhibited by ACE inhibitors [24][25][26]31,32]. According to Fleming [33], chymase functions as a tissue Ang II-forming enzyme in human heart, arteries, and lungs while ACE accounts for the majority of Ang II formation from Ang I in the circulation. Chymase broad functional activities includes its participation in collagen IV turnover [34], hydrolysis of apolipoprotein E [35], thrombin, and fibrin [36,37].
Although numerous studies confirmed that chymostatin is a specific chymase inhibitor [4,5,13,16,20,28,29,32,[38][39][40][41][42][43][44][45], this inhibitor may also nonspecifically reduces the activity of cathepsin G, cysteine proteases, and high-molecular weight proteasomes [46]. A recent report suggested that elastase-2, primarily expressed in human and mouse epidermis may represent an alternate pathway for Ang II generation from Ang I in conditions of chronic ACE inhibition in rats but not human vascular tissue [30]. While our past studies did not exclude a potential contribution of a chymostatin-sensitive elastase-2 to Ang-(1-12) metabolism in the rat [13], our current findings in human atrial myocytes suggest that human elastase does not account for our findings. This interpretation is based in part on the demonstration that chymostatin inhibited the hydrolysis of the colorimetric N-suc-AAPF-pNA substrate by human plasma membranes. Although Nsuc-AAPF-pNA is not hydrolyzed by human leukocyte elastase [47], this finding did not exclude a contribution of human pancreatic elastase. Moreover, our additional finding that the orally active chymase inhibitor (TEI-F00806) [16] prevented the hydrolysis of Ang-(1-12) by 95% strengthens the importance of cardiac chymase as the primary enzyme accounting for Ang-(1-12) metabolism in human samples. This interpretation is further reinforced by the concurrent measurements of chymase protein, a strong correlation between chymase activity and expression, and the detection of chymase immunoreactive products in the atrial human tissue.
While studies of Ang-(1-12) processing in cardiac myocytes from neonatal rats showed a contribution of ACE and neprilysin to Ang-(1-12) metabolism in WKY, an increase in cardiac chymase activity in Ang-(1-12) metabolism was found in SHR [13]. Prosser et al. [4,5] reported that both ACE and chymase contributed to Ang-(1-12) metabolism in the vascular system and the heart of Sprague Dawley rats. Since the Ang II generating capacity from Ang-(1-12) by ACE appeared to be predominantly found in the vasculature, these data suggest tissue specific differential roles for ACE and chymase in Ang-(1-12) metabolism.

Ethics Statement
The study has been approved by the Wake Forest University Medical Center (IRB# 00004355; approved on November 30th 2010). The human samples were collected from the patients undergoing cardiac surgery (Table 2)

Human Tissue (atrial appendage)
Human left atrial appendage tissue from 5 men (Mean 6 SD, 59613 yrs.) and 4 women (Mean 6 SD, 7563 yrs.) were obtained from patients undergoing cardiac surgery at Wake Forest University Baptist Medical Center (Winston-Salem, NC). All patients underwent surgical excision of the left atrial appendage (LAA) as a routine part of the MAZE surgical procedure [51]. Diagnosis, drug treatment and clinical status of all nine patients are described in Table 2. The excised LAA tissue obtained from these patients was kept in cold cardioplegic solution from the time of removal, frozen within 15-30 minutes, and stored at 280uC until metabolism studies. The use of these tissues was approved by the Wake Forest University Medical Center Institutional Review Board.

Histology and Immunohistochemistry
Left atrial tissues from human cardiac patient were fixed in 4% paraformaldehyde for 24 h and then transferred into 70% ethanol. After dehydration, the tissues were embedded in paraffin and each section was cut in 5-mm thick sections. Immunohistochemistry was performed using well characterized protein A purified polyclonal antibody directed to the COOH-terminus of full length of human  Plasma membrane preparation Membranes were prepared at 4uC in a manner similar to the approaches previously described by our laboratory [52] and by Urata et al. [26] Frozen atrial tissue (50-100 mg) was homogenized at 4uC in 1 mL PBS (pH 7.4) in a Qiagen TissueLyzer (Valencia, CA). The homogenate was centrifuged at low spin (200 g) for 5 min at 4uC to remove the connective tissues and cell debris. The supernatant was transferred into a new tube and centrifuged at 28,000 g for 20 min at 4uC to collect the plasma membrane. The pellet was resuspended in PBS and centrifuged as described above. Finally, the pellet was resuspended in PBS and stored at 280uC till its use for metabolism studies. The concentration of each cardiac membrane preparation is expressed in terms of protein concentration measured by Bradford Reagent using bovine serum albumin (BSA) as the standard; the results were normalized in terms of per mg protein.

I-Ang-(1-12) Metabolism by Human Plasma Membranes
Metabolism of human Ang-(1-12) by plasma membrane isolated form atrial tissue was studied under different combinations of RAS inhibitors as documented in Table 3.
For 125 I-Ang-(1-12) metabolism studies, plasma membranes (50 mg per reaction mixture) were preincubated for 15 min under various combinations of RAS and peptidases inhibitors (50 mM each). After preincubation of plasma membranes with the inhibitors, human 125 I-Ang-(1-12) (1 nmol/L; specific activity 3,900 cpm/fmol) was added to the reaction medium and incubated for an additional 60 min at 37uC. At the end of the incubation time, the reaction was stopped by adding equal volume of 1% phosphoric acid, mixed well, and centrifuged at 28,000 g for 20 min to remove the plasma membrane. The clear supernatants were stored at 220uC until processing the samples for Ang contents [Ang- (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12), Ang I, Ang II and Ang-(1-7)] by HPLC analysis. On the day of HPLC analysis, the samples were filtered before separation by reverse-phase high-performance liquid chromatography. We used a linear gradient from 10% to 50% mobile phase B at a flow rate of 0.35 mL/min at ambient temperature. The solvent system consisted of 0.1% phosphoric acid (mobile phase A) and 80% acetonitriles/0.1% phosphoric acid (mobile phase B). The eluted 125 I products were monitored by an in-line flow-through gamma detector (BioScan Inc., Washington, DC). Products were identified by comparison of retention time of synthetic [ 125 I] standard peptides and the data were analyzed with Shimadzu LCSolution (Kyoto, Japan) acquisition software. The iodination of human Ang-(1-12) and other angiotensins were performed as described previously [53].

Contribution of Specific Enzyme (ACE or Chymase)
The contribution of specific enzyme (ACE or chymase) to 125 I-Ang-(1-12) hydrolysis were analyzed by measuring the amount of Ang products generated after exposing 125 I-Ang-(1-12) with plasma membrane in the presence of all RAS inhibitors cocktail and in the absence of specific enzyme inhibitors for ACE or chymase only (as described above). Supernatants were collected after exposing the 125 I-Ang- (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12) in the presence of the plasma membrane for 60 min at 37uC under two different enzyme inhibitors conditions (+All RAS inhibitors versus minus ACE/ chymase inhibitor only) and were analyzed by HPLC as described above. The enzyme activity was calculated based on adding 1 nmol/L of 125 I-Ang-(1-12) substrate to the reaction mixture and determining the amount of Ang II product formation. Experiments were performed three or more times and the enzyme activity values were reported as fmoles of Ang II product formation from 125 I-Ang-(1-12) substrate per min per mg protein (fmol Ang II formation6min 21 6mg 21 ).

Western blot Analysis of Human Chymase
The expression of chymase protein in human plasma membrane samples were assessed by immunoblot technique as previously described [54]. Briefly, the plasma membranes (50 mg protein) were separated by gel electrophoresis (10% gel) and transferred to polyvinylidene defluoridated membranes (PVDF). The PVDF membranes were probed with a primary monoclonal anti-human chymase antibody (CMA1 antibody from R&D System, Minneapolis, MN, Cat# MAB4099; 2 mg/mL) and mouse anti-b-actin (1:5,000; Sigma-Aldrich, St. Louis, MO, Cat# A-5441). After incubation with the primary antibody, the membranes were probed with the horseradish peroxidaseconjugated secondary antibody (anti-mouse, 1:5,000; Pierce Inc., Rockford, IL, USA). Immune complexes were visualized using ECL plus detection reagents (Pierce). Densitometric analysis was performed by measuring the intensity of all Western Blotting bands with the MCID imaging system (MCID Elite 7.0, Imaging Research Inc., St. Catharines, ON, Canada).

Statistical Analysis
Experiments were repeated independently three or more times. All values are reported as means 6 SEM. The Student's t-test and repeated-measures ANOVA followed by a Turkey's post hoc test for multiple comparisons were used to determine significant differences at P,0.05 using GraphPad Prism 5.0 software (San Diego, CA). Table 3. Outline of enzyme inhibitors employed in the experiments.

Group
Inhibitors added (50 mM each)

Minus RAS inhibitor groups
All above inhibitors except one of the RAS inhibitor (lisinopril or chymostatin) was omitted at a time from the reaction mixture. doi:10.1371/journal.pone.0028501.t003