|Home | About | Journals | Submit | Contact Us | Français|
The beneficial effects of angiotensin I-converting enzyme (ACE) inhibitors go beyond the inhibition of ACE to decrease angiotensin (Ang) II or increase kinin levels. ACE inhibitors also affect kinin B1 and B2 receptor (B1R and B2R) signaling, which may underlie some of their therapeutic usefulness. They can indirectly potentiate bradykinin (BK) and ACE-resistant BK analogues' actions on B2Rs to elevate arachidonic acid and NO release in laboratory experiments. Studies indicate that ACE inhibitors and some Ang metabolites increase B2R functions as allosteric enhancers by inducing a conformational change in ACE. This is transmitted to B2Rs via heterodimerization with ACE on the plasma membrane of cells. ACE inhibitors are also agonists of the B1R, at a Zn-binding sequence on the second extracellular loop that differs from the orthosteric binding site of the des-Arg-kinin peptide ligands. Thus, ACE inhibitors act as direct allosteric B1R agonists. When ACE inhibitors enhance B2R and B1R signaling, they augment NO production. Enhancement of B2R signaling activates eNOS yielding a short burst of NO; activation of B1Rs results in a prolonged high output of NO by iNOS. These actions, outside inhibiting peptide hydrolysis, may contribute to the pleiotropic therapeutic effects of ACE inhibitors in various cardiovascular disorders.
Millions of patients are treated with angiotensin 1-converting enzyme (ACE) inhibitors to combat hypertension, congestive heart failure or diabetic renal diseases 1-4. ACE inhibitors significantly reduce mortality after myocardial infarction 3 and are beneficial in other high risk patients.
ACE inhibitors block the metabolism of several peptides by ACE, notably the conversion of angiotensin (Ang) I to II 5, and the inactivation of bradykinin (BK) 6-8 or the hemoregulatory tetrapeptide Ac-Ser-Asp-Lys-Pro 9. The conversion of Ang I to Ang II was first found to occur in horse plasma 5, and one of us reported the identity of ACE with kidney and human plasma kininase II we discovered before 6-8, 10, 11. Consequently, a single peptidyl-dipeptidase not only releases the hypertensive Ang II, but also inactivates the hypotensive BK. How much of the therapeutic effectiveness of ACE inhibitors is due to blocking Ang II release 12 or to prolonging the short half-life of BK 7 and its congener Lys1-BK (kallidin) has been debated. This is further complicated by the existence of two kinin receptors. The first characterized, but incongruously named, B2 receptor (B2R) is activated by native BK or kallidin 13. The second, so-called B1 receptor (B1R), does not bind native kinins; its ligands are metabolites of BK and kallidin lacking the C-terminal arginine 14 removed by plasma carboxypeptidase (CP) N 15-17 or membrane CPM 18-20. Whereas the B2R is widely expressed constitutively, B1R expression is usually induced after noxious stimuli or by inflammatory cytokines 13, 14, 21-23, although some cells (bovine lung endothelial or human fibroblasts) express B1Rs constitutively.
ACE inhibitors can enhance both B2 and B1R signaling. Blocking kinin inactivation by ACE raises the concentration of intact B2R agonists, which are also the substrates of CP N and M. This can generate more des-Arg-kinin B1R agonists (Fig. 1). The successful use of antagonists of the Ang II AT1 receptor (AT1R) for many of the same indications as ACE inhibitors, does not prove ACE inhibitors work only through reducing Ang II as there are complex interrelationships among Ang II, BK and their receptors. Ang II has two receptors, AT1R and AT2R. AT1R is blocked by drugs such as losartan, which can shift Ang II actions to the AT2R. This switching of receptors further counteracts AT1R effects as it leads to the release of mediators such as nitric oxide (NO), and is attributed partially to release of BK to activate B2Rs, a form of “crosstalk” 12, 24, 25. Intricate Ang II and kinin receptor interrelationships were also indicated in animal experiments where both kinin B1 and B2Rs share in the favorable cardiovascular effects of AT1R blockade 26, 27.
Despite the established role of Ang and kinin receptors in ACE inhibitor effects, much remains puzzling. For example, BK levels in plasma, even after ACE inhibitors, are lower than the effective doses of exogenous peptide. Crosstalk between AT2Rs and B2Rs does not necessarily require the release of BK by kallikrein, because kallikrein and other proteases can directly activate the B2R 28-31. Suggestions that ACE inhibitors lead to an “accumulation” of BK are not plausible.
We suggest that ACE inhibitors have actions that go beyond blocking covalent peptide bond hydrolysis and may explain some of their therapeutic effectiveness. As summarized below, we propose that these drugs are also allosteric effectors of B1Rs and B2Rs, functioning as direct allosteric agonists of B1Rs and as indirect allosteric enhancers of kinin activity on B2Rs via interactions with ACE.
Regarding allosterism, Monod et al 32 concluded that in regulatory proteins, indirect interactions between distinct specific binding sites explain their regulatory function. Monod and Jacob's presentation at a meeting in 1961 was considered revolutionary 33. Proteins could recognize more than one molecular partner at an unrelated second site, forming the basis for allosteric effects 32. Put another way, the two sites can “talk” to each other and, as we understand it, this “talking” is mediated by conformational change. According to the International Union of Pharmacology Committee on Receptor Nomenclature 34, allosteric enhancers are modulators of receptor function that enhance orthosteric ligand affinity and/or agonist efficacy, while having no effect on their own. They can indirectly bias receptor signaling to endogenous agonists by allosteric modification of the receptor 35, 36. G protein-coupled receptors (GPCRs), such as the B2R, are prototypes of allosteric proteins that can exist in many different conformational states. Allosteric transition involves isomerization and stabilization of one of the receptor's many conformational states, each of which may have a different affinity for a ligand and/or preferentially activate different signal transduction pathways. Thus, allosteric enhancers may affect only a subset of a receptor's full signaling pathway; this “collateral efficacy” can lead to new drug discovery 35.
Investigations into allosteric modulators of GPCRs have focused on small molecule ligands 37, but we propose that ACE serves this function with the B2R. Thus, ACE inhibitors enhance mediator release (e.g., NO, prostaglandins or EDHF) 13, 38 not only by protecting BK from degradation, but also indirectly by binding to ACE, which allosterically enhances B2R activation by orthosteric ligand BK (Fig. 1).
The first clinically tested ACE inhibitor, teprotide, came from snake venom peptides that potentiated BK action on isolated organs 39. Although the Bothrops jararaca venom peptides BPP5a and BPP9a inhibited ACE 39, with other BK potentiating peptides or analogues, ACE inhibition, kinin inactivation and BK potentiation did not correlate 40. This indicate an additional mechanism(s). For example, adding an ACE inhibitor to the isolated guinea pig ileum when isotonic contraction to BK was at its maximum, rapidly doubled its magnitude 41. This was not due to blocking BK inactivation as its t1/2 was 12-15 min in the ileal preparation. On guinea pig atria, enalaprilat enhanced the positive inotropic effect of BK, but not by blocking inactivation 42. Investigators, using different techniques, found that ACE inhibitors potentiated BK actions in tissue preparations not due to inhibition of BK degradation. This led to the supposition that ACE inhibitors acted on B2Rs or stimulated a crosstalk between ACE and the B2R 42-47.
Experiments with cultured cells led us to conclude that ACE inhibitors induce a conformational change in ACE that is transmitted to the B2R owing to their close contact, resulting in potentiation or reactivation of receptor signaling. For example, enalaprilat preserved the B2R in high affinity form that increased arachidonic acid release 46. The inhibitor also resensitized B2Rs, which had been desensitized to BK, and reduced its internalization. B2Rs are desensitized after prolonged exposure to kinins and become unresponsive to a second dose of ligand. Thus, resensitization by ACE inhibitors causes the B2R to react again to agonist already present in the medium. The resensitization of B2R by ACE inhibitor was confirmed with porcine endothelial cells and attributed to blocking sequestration of receptor into caveolin-rich membranes 45. These and numerous additional experiments indicate that ACE inhibitors can act through ACE as indirect allosteric enhancers of BK effects on B2R, resulting in increased mediator release 46-53.
More proof of these indirect effects of ACE inhibitors on B2R signaling was obtained with BK analogues such as HT-BK (~50% resistant to cleavage by ACE) 41, 42, 46, 48 or other almost completely resistant BK analogues 49-53. Generally, ACE only cleaves oligopeptides of less than 13 residues efficiently 46, 50, 52, but B2R agonists can be larger molecules. For example, BK coupled at the N-terminus to soluble dextran 54 or Lys1-BK dansylated at the α and ε amino groups of Lys1, remained B2R agonists not appreciably cleaved by ACE 52. Another ACE-resistant B2R agonist has a non-peptide bond [Phe8ψ(CH2-NH)Arg9] at the C-terminus 55. Even with these BK analogues, ACE inhibitors still augmented B2R signaling obviously without preventing enzymatic degradation.
ACE inhibitor enhancement of B2R action involves signaling pathways different from those stimulated by B2R agonists alone. For example, B2R activation by BK is not affected by protein kinase C and phosphatase inhibitors, but they blocked the resenzitization of the B2R to BK by an ACE inhibitor 49, 52. ACE inhibitors also decreased B2R phosphorylation 47, 49, suggesting the involvement of phosphorylation and dephosphorylation of the B2R in this response. The tyrosine kinase inhibitor genistein also blocked B2R resensitization caused by ACE inhibitors or Ang1-7, and Ang1-9 peptides 52, indicating that these endogenous peptides are also allosteric enhancers of B2R function (see below).
For ACE inhibitors to augment BK responses, ACE and the B2R must be coexpressed on the plasma membrane, close enough to transmit allosteric effects. For example, ACE inhibitors did not potentiate BK effects in cells lacking ACE or when purified soluble ACE was added to the medium 56. Furthermore, B2R expression can modulate ACE activity 57. We used several techniques to show that ACE and B2Rs interact 56. ACE and the B2R co-immunoprecipitated and confocal microscopy revealed that they were co-localized on the membrane. Using B2Rs tagged at C-terminus with yellow fluorescent protein (YFP) (acceptor) and ACE labeled at the C-terminus with donor cyan fluorescent protein (CFP), fluorescence resonance energy transfer (FRET) indicated that the fluorophores attached to ACE and B2R were within 10 nm on the membrane 56.
The ACE molecule resides primarily extracellularly, (ratio of extracellular to transmembrane and cytosolic portions ~ 26:1), quite different from the human B2R, where the ratio of extracellular amino acids to the transmembrane and cytosolic portions is ~1:11. To establish whether the extracellular domains are also close enough for FRET, we labeled B2R with an N-terminal YFP and ACE with an N-terminal CFP and detected significant FRET, indicating the extracellular portions are also within 10 nm (Chen et al., to be published).
The formation of ACE/B2R heterodimer on membrane should be a bimolecular reaction dependent on reactant concentrations. If ACE is in excess, ACE inhibitors could more effectively enhance the activation of B2R by kinins, whereas if cells express many more B2Rs than ACE, ACE inhibitors would not be effective 56.
The precise interaction sites and orientation of B2Rs and ACE on the membrane are still unknown, but biochemical and structural features provide clues. Somatic ACE has two active sites contained in N- and C-domains, connected by a bridge section 58. The short transmembrane anchor is followed by a cytosolic tail 59 (Fig. 1). The cytosolic and transmembrane domains of ACE are not required for interaction and potentiation of B2R responses 53. However, for ACE to interact with the B2Rs, ACE has to be membrane anchored for proper orientation. A chimeric B2R molecule with ACE fused to its N-terminus had ACE activity and was a functional B2R, but ACE inhibitors did not potentiate B2R responses 56. Finally, with an ACE construct containing the C-domain active site but lacking most of the N-domain, ACE inhibitor still potentiated B2R response 60. This suggests that the extracellular C-domain interacts with B2Rs (Fig. 1). This is consistent with the ACE crystal structure and modeling indicating 2 possible orientations for N- and C-domains 61, placing the N-domain either ~36 Å or 72 Å from the membrane. Based on the crystal structure of the β-adrenergic receptor 62, extracellular domains of the B2R could extend maximally ~20 Å above the membrane.
Supportive evidence indicates that binding of ACE inhibitors alters the conformation of ACE, a necessary movement for allosteric modification of the B2R. For example, ACE inhibitors induce phosphorylation of ACE's C-terminal tail, thereby activating signal transduction pathways to increase expression of proteins such as COX-2 63. The two active domains of ACE have different specificities64 and exhibit negative cooperativity 65; binding of ACE inhibitor to one domain alters the conformation of the second domain, rendering it inaccessible to a second inhibitor. The crystal structures of the C- or N- domain revealed a deep active site cleft closed to exterior by a “lid”; consequently, a conformational change would be required for the access of substrate or inhibitor to the enzyme 61, 66, 67. Modeling of the ACE structure indicates intrinsic flexibility around the active site, suggesting a hinge mechanism to open it for substrate/inhibitor binding 67. Thus, inhibitor binding to ACE does change its conformation, which can be transmitted to the B2R owing to their close association on membrane 56, resulting in allosteric enhancement of B2R signaling (Fig. 1).
ACE inhibitors affect the two kinin receptors differently 21-23, 46, 47, 50. The human B1R is directly activated by ACE inhibitors, even in absence of ACE expression 21-23, 68 (Fig. 1). ACE inhibitors activate B1Rs to release NO, most consistently in cultured human endothelial cells via iNOS, in the same range as the B1R ligands, desArg9-Lys1-BK or desArg9-BK (10-9 - 10-6 M) 21-23. Although desArg9-Lys1-BK has higher affinity for the B1R than desArg9-BK, the two are about equally active on human B1Rs 23 owing to their efficacies. Enalaprilat, quinalaprilat, ramiprilat, or captopril are active B1R agonists, while lisinopril is not at the same concentrations, probably due to its positively charged ε-NH2 22.
In ACE, the canonical pentamer sequence (HEXXH) containing zinc-binding residues in the N- and C-domain active sites, are important for inhibitor binding 59. In metallopeptidases, all three zinc-binding residues (2 His and 1 Glu) don't have to be sequential provided they are sterically close (within about 3 Å) in the 3-D structure. The second extracellular loop of the human B1R has the same consensus sequence (HEAWH; residues 195-199) required for ACE inhibitor (but not peptide ligands) to activate B1Rs 22 (Fig. 1). Thus, allosteric ACE inhibitor activity is blocked by agents or mutations not affecting orthosteric des-Arg-kinin activity. For example, a synthetic undecapeptide containing the pentamer (LLPHEAWHFAR; residues 192-202), blocked B1R activation by enalaprilat but not by des-Arg-kinin 22, 23. Mutation of H195 to Ala in human B1R did not affect peptide agonist action, but enalaprilat's effect was much reduced 22 (and Tan, et al, to be published).
B1R activation can increase inflammation, pain and fibrosis in diabetic cardiomyopathy 13, 14, 69, but it is also beneficial after myocardial infarction in rats or mice 27, 70, 71. Increased NO synthesis, owing to B1R activation 21, 72, may also contribute to ACE inhibitors' therapeutic effects after an MI, and protect cardiomyocytes 73. NO release, after ACE inhibitor activation of B1R, inhibited protein kinase Cε (PKCε) 23 that can benefit the failing heart 74. B1R signaling was recently reported to prevent homing of encephalitogenic T-lymphocytes into the CNS, which was enhanced in B1R-/- mice 75. CPM, closely associated with myelin centrally and peripherally 76, should contribute by generating B1R ligands. The report mentioned that ACE inhibitor also suppresses inflammation in the CNS 75.
Without carboxypeptidases, endogenous orthosteric B1R ligands could not be generated and B1R signaling would not occur. CPM and B1Rs interact on the cell membrane 77 and based on CPM's crystal structure and modeling 20, its active site would be properly oriented along the membrane to deliver agonist effectively to B1R. In bovine or human endothelial cells, B2R agonists cause B1R-dependent release of calcium or generation of NO 77, 78, which also depended on CPM.
Activation of B1 and B2Rs can promote inflammation or intensify pain 13, 14 but can also improve the functions of the failing heart or kidney 4, 12, 13, 26, 27, 70, 79. B1 and B2Rs both activate NO synthesis, but B2R agonists stimulate transient eNOS-derived NO whereas B1R activation leads to prolonged high output NO via iNOS 21, 22, 72. ACE inhibitors do not activate B1Rs in blood vessels lacking endothelium, where peptide ligands are vasoconstrictor 14.
ACE inhibitors can potentiate kallikrein-mediated stimulation of B2Rs, independent of kinin release 29, 30, but after prekallikrein activation 80. Plasma prekallikrein may also be allosterically activated by prolylcarboxypeptidase 81 or heat shock protein 90 82. This could result from induction of a conformational change in prekallikrein, exposing it to another protease or to trace autocatalytic activity, yielding activated kallikrein 83, 84.
Endogenous peptides, such as Ang derivatives Ang1-7 and Ang1-9, can also augment orthosteric BK effect on B2R 52, 85. Ang1-9 is released from Ang I by a carboxypeptidase 86 or by cathepsin A (deamidase) 85, 87, 88. Ang1-9, a relatively stable intermediate, is also liberated by human heart tissue 85, 88. Ang1-7 is cleaved from Ang I by human neprilysin 89 and from Ang II by ACE2 90, 91 and prolylcarboxypeptidase 92. Ang1-7 counteracts Ang II actions for example by improving baroreceptor reflex and decreasing vascular and smooth muscle growth. Ang1-7 activates the Mas receptor and also potentiates BK effects in vivo 91. Both Ang1-9 and Ang1-7 can inhibit ACE, but they augment BK effects on B2Rs at orders of magnitude lower concentrations in cultured cells than their IC50 values 52, 85. Thus, Ang1-7 and Ang1-9 could antagonize Ang II effects in vivo, also as allosteric enhancers of the B2R.
We did not, and could not, aim to complete the history of ACE inhibitors leaving no major questions unanswered, but sought to summarize some modes of actions that may contribute to the efficacy of these drugs. The complexities make it difficult to interpret their effects as due only to a single mediator. ACE cleaves other active peptides besides Ang I and BK and ACE inhibitors enhance responses of kinin receptors beyond blocking kinin catabolism 29, 46, 93, 94. Exogenous ACE inhibitors and endogenous Ang1-7 and Ang1-9 peptides are indirect allosteric enhancers of B2R activation by the orthosteric peptide ligands. They augment collateral efficacy by inducing conformation changes via ACE and B2R complexes on cell plasma membranes. This leads to enhanced release of mediators such as NO, EDHF 38 or prostaglandins13. ACE inhibitors are also direct activators of B1Rs at an allosteric site that differs from the orthosteric site of peptide ligands. The consequence is a prolonged high output NO production by iNOS in human endothelial cells 22, 23, 72. Finally, ACE inhibitors can potentiate direct actions of kallikrein on the B2R in the absence of kinin release 29, 30, 95.
Sources of Funding: This work was supported by NIH Grants HL60678, DK 41431 and HL36473.
Conflict of Interest: None