ACE in renal development.

Studies in both mice and humans have revealed a role of ACE in kidney development^16–18. Global Ace^−/− mice are viable, probably because some renal development in mice takes place after birth^16,17. Although blood pressure is profoundly reduced in these mice (a reduction of systolic blood pressure from approximately 110 mmHg to 73 mmHg)¹⁷, the kidney structure is identical to that of wild-type (WT) mice at birth. However, by post-natal day 16, Ace^−/− mice exhibit expansion of the renal pelvis and underdevelopment of the renal medulla¹⁹. An equivalent phenotype is observed in mice with a genetic knockout of any of the RAS components (that is, angiotensinogen, renin or all type 1 angiotensin II receptors (AT1Rs))^20–22. Mechanistically, angiotensin II stimulates ureteral smooth muscle growth and uni directional peristaltic waves that originate in the renal pelvis and assist urine transport along the ureters to the bladder²³. The absence of such waves in ACE-deficient mice (or any mouse model lacking either angiotensin II or AT1Rs) results in inefficient urine transit that elevates pressures within the renal pelvis and ureter. Hence, the hydronephrosis that occurs in these mice is due to a functional defect and not physical blockage of the outflow tract²³. Taken together, ACE and the RAS facilitate the efficient movement of urine away from the kidney and the normal development of the renal outflow tract.

In contrast to mice, human kidney development primarily occurs in utero. Genetic mutations in ACE, or any functional interruption of the RAS, can result in renal tubular dysgenesis (RTD). One study of RTD reported a series of 54 mutations in genes encoding RAS proteins in 48 families, of which ACE mutations were reported in 31 families¹⁸. Absence of a functional RAS results in severe hypotension in the fetus, leading to renal hypoperfusion, RTD and oligohydramnios, which cause fetal compression. Severe renal insufficiency and lung hypoplasia usually cause infants to die rapidly after birth.

ACE and granuloma.

ACE is upregulated in several diseases characterized by granulomas, including active sarcoidosis, histoplasmosis, leprosy, silicosis, granulomatosis with polyangiitis and the granulomatous reaction induced by murine schistosomiasis^44–48. Whether the formation of a physical barrier is advantageous or disadvantageous to the host is currently unclear⁴⁹. On the one hand, the physical isolation (‘walling-off’) of bacteria may reduce the rate of disease dissemination⁵⁰. On the other hand, as oxygen is a major component of antimicrobial activity, reduced access to oxygen for myeloid cells within the granuloma could be a potential disadvantage, despite these cells being in close proximity to the invading pathogen. Although granulomas were first described in 1679, true mechanistic insights into the basic biochemical signals that induce granuloma formation have only emerged in the past 15 years⁵¹. Granulomas are composed of mature macrophages that form in response to a chronic stimulus (infection or foreign body). The cytokine TNF seems crucial in this process^52,53. These activated macrophages are larger with more cytoplasm and intracellular organelles and a more ruffled membrane than mononuclear phagocytes in blood^50,54. Within a granuloma, macrophages can fuse into giant cells, perhaps in response to IL-4 (REF. 55). Macrophages within the granuloma were reported to have an approximately sevenfold increase in ACE mRNA compared with kidney macrophages in a zebrafish model of Mycobacterium marinum infection, in which detailed cellular events that characterize granuloma formation can be visualized⁵⁶. Fluorescence imaging and electron microscopy showed that forming granulomas undergo macrophage reprogram ming to express adherens-type cellular proteins, such as epithelial cadherin (E-cadherin; also known as CDH1), which are more typical of epithelial cells than myeloid cells, thereby allowing macrophages to form cell–cell contacts. The biochemical signals that induce the ‘epithelial reprogram ming’ described in zebrafish granuloma may also contribute to the observed increase in ACE expression.

In summary, increased ACE expression by myeloid cells within a granuloma has been known for many years. However, why elevated macrophage ACE expression is such a consistent biochemical feature of a granuloma is currently unclear. One possibility is that ACE increases the antibacterial effectiveness of macrophages, as discussed below.

ACE and macrophage function.

Macrophages are important in the initial innate immune response and also have a crucial role as antigen-presenting cells (APCs) by interacting with T cells in the adaptive response. Mice overexpressing ACE in myeloid-derived cells only, known as ACE 10/10 mice, were initially developed to study the role of angiotensin II in the development of atherosclerosis but have proved useful in advancing our understanding of how ACE overexpression can affect innate and adaptive immunity⁵⁷. Macrophages from ACE 10/10 mice produce 16-fold to 25-fold more ACE protein than do cells from WT littermates (FIG. 2a), whereas neutrophils and dendritic cells mildly overex-press ACE, and expression in T cells and B cells is similar to that of WT cells. By contrast, ACE expression is absent in endothelial cells, which are the main source of ACE in WT mice. Circulating ACE levels in ACE 10/10 mice are similar to those in WT mice and ACE 10/10 mice have a normal basal blood pressure.

Open in a separate window

Figure 2 |

Macrophage-specific ACE overexpression suppresses tumour growth.

a | ACE 10/10 mice overexpress angiotensin-converting enzyme (ACE) in macrophages. ACE is predominantly located on the cell surface but is also present within the endoplasmic reticulum (ER) and probably within endosomes. b | Analysis of tumour volume 14 days after intradermal implantation of B16-F10 melanoma cells into ACE 10/10 mice, ACE 10/10 mice crossed with ACE wild-type (WT) mice (heterozygous (HZ) mice) and WT mice shows that ACE overexpression in macrophages attenuates tumour growth. Data from individual mice (open blue diamonds) as well as the group means (filled orange circles) and standard error of the mean are shown. c | Representative images of the tumours at day 14 in ACE 10/10 mice and ACE WT mice are presented. Parts b and c are adapted with permission from REF. 57, Elsevier.

Insights into the functional relevance of ACE expression by myeloid cells came from a study of tumour development, in which tumours arising from implantation of B16 melanoma in ACE 10/10 mice were on average only 17% the size of those in WT mice 2 weeks after implantation⁵⁷ (FIG. 2b,c). Similar differences in tumour size were also observed following implantation of tumour cells in ACE10/10 mice on an outbred genetic background and following intravenous injection of tumour cells^57,58. The increased immune response originates from bone marrow-derived cells, as WT recipients of bone marrow from ACE 10/10 mice developed melanoma tumours only 51% the size of those in WT mice receiving bone marrow from WT mice. Importantly, the difference in B16 melanoma tumour size between ACE 10/10 mice and WT mice was abrogated by treatment with the ACE inhibitor captopril. By contrast, treatment of ACE 10/10 mice with the angiotensin receptor blocker losartan did not alter tumour size compared with that of untreated ACE 10/10 mice, indicating that the catalytic activity of ACE is neces sary to inhibit tumour growth but that this effect is not mediated by the effects of angiotensin II on AT1 (REF. 57). Consistent with this finding, when challenged with B16 melanoma, angiotensinogen-deficient mice with ACE-overexpressing myeloid cells develop tumours less than one-third the size of those in mice with WT levels of ACE expression in myeloid cells (Supplementary Figure S1). Taken together, these data suggest that tumour resistance in ACE 10/10 mice is not due to the absence of endothelial ACE but is rather due to overexpression of catalytically active ACE by monocytes and macrophages independent of angiotensin II signalling, thereby potentially facilitating an enhanced immune response⁵⁷.

In support of this hypothesis, ACE 10/10 macro phages increase the production of pro-inflammatory cytokines — such as IL-12β, TNF or nitric oxide — beyond WT levels in response to lipopolysaccharide (LPS) or IFNγ in vitro or to B16 melanoma in vivo^4,57.

To evaluate the innate immune response and test the physiological relevance of the B16 melanoma model, we challenged ACE 10/10 mice with either Listeria monocytogenes or with methicillin-resistant Staphylococcus aureus (MRSA)⁵⁹. Five days after infection with L. monocytogenes, the spleens and livers of ACE 10/10 mice had 8.0-fold and 5.2-fold less bacteria than WT mice, respectively. MRSA was placed subcutaneously in the skin and, similar to observations in the melanoma model, skin lesions in these mice after 4 days of inoculation were on average 22% the size of those in WT mice and contained ~50-fold fewer viable bacteria than the lesions of WT mice. As in the B16 melanoma model, treatment of ACE 10/10 mice with the ACE inhibitor lisinopril restored the WT phenotype. Thus, ACE 10/10 mice have an enhanced immune response beyond that of WT mice in models of both innate and adaptive immunity.

ACE and antigen processing.

Major histocompatibility complex (MHC) class I antigens have a crucial role in the defence of an organism against intracellular pathogens and perhaps also against tumours. All nucleated cells can present fragments of intracellular proteins on the plasma membrane via the MHC class I complex. Cytoplasmic peptides are imported into the endoplasmic reticulum (ER) through peptide transporter involved in antigen processing (TAP)^60,61. In the ER, MHC class I peptides are trimmed and loaded onto MHC class I proteins, from where they undergo vesicular trafficking to the plasma membrane (FIG. 3). These MHC class I-bound peptides can be recognized by CD8⁺ T cells either as self-peptides or as non-self-peptides (for example, of viral origin), thereby eliciting CD8⁺ T cell inactivation and activation, respectively. In addition, specialized APCs, such as macrophages and dendritic cells, take up cellular debris and efficiently present MHC class I protein-bound viral peptides to CD8⁺ T cells.

Open in a separate window

Figure 3 |

ACE participates in peptide trimming during antigen processing.

Angiotensin-converting enzyme (ACE) overexpression in antigen-presenting cells (APCs) likely changes the display of both major histocompatibility (MHC) class I and MHC class II epitopes. MHC class I and MHC class II proteins bind and display peptides on the surface of cells. As a peptidase, ACE can cut peptides and thereby affect the diversity of peptides that are presented to T cells. a | Cytoplasmic proteins are proteolytically cleaved by the proteasome to become potential MHC class I peptides, which are imported into the endoplasmic reticulum (ER) via peptide transporter involved in antigen processing (TAP). ACE helps trim MHC class I peptides within the ER, which occurs in both wild-type and ACE-overexpressing APCs. b | ACE expression also changes MHC class II presentation of peptide epitopes, although the exact subcellular location and mechanism is somewhat unclear. CLIP, class II-associated invariant chain peptide; Ii, invariant chain; MVB, multivesicular body; TCR, T cell receptor.

Evidence from bone marrow transplantation and other experiments in mice suggests that ACE enzymatically trims both the self and non-self MHC class I peptides before they are bound to MHC class I proteins and displayed by cells^62,63 (FIG. 3a). Donor splenocytes from ACE-deficient mice were recognized as foreign by CD8⁺ T cells in WT syngeneic mice, and vice versa, splenocytes from WT donor mice expressing ACE were also recognized as non-self in ACE-deficient mice. These findings imply that ACE-deficient APCs process peptides differently from those that express ACE.

ACE is normally a cell surface protein and traffics through the ER. To examine whether ACE is active in the ER, we used an approach that functionally isolated the ER from the cell cytoplasm by using a cell line lacking TAP, which is thus unable to import cytoplasmic peptides into the ER⁶³. The cell line was modified to express a short peptide from chicken ovalbumin (OVA) with a signal sequence that allowed it to enter the ER in the absence of TAP and without passing through the cytoplasm. When comparing the MHC class I peptide-processing ability of truncated ACE — which lacks the signal peptide and thereby does not enter the ER — with that of full-length ACE, which does enter the ER, we found that carboxy-terminal peptide cleavage of the OVA peptide within the ER only occurred in cells with full-length ACE. These data suggest that ACE is catalytically active in the ER, a conclusion consistent with in vivo analysis⁶³. Additional experiments using purified cellular ER may serve to further support this conclusion.

Moreover, ACE 10/10 mice — when challenged with B16 melanoma cells constitutively expressing OVA — develop elevated levels of CD8⁺ T cells directed against both an intrinsic B16 melanoma MHC class I epitope and an OVA MHC class I epitope⁵⁷. Thus, ACE over-expression in myeloid cells facilitates a greater CD8⁺ T cell response than do WT levels of ACE expression in myeloid cells. This finding is particularly interesting, given that the immune pathway(s) by which myeloid ACE overexpression increases the immune response are currently unknown. Taken together, however, these findings strongly suggest that ACE plays a crucial role in peptide antigen processing and presentation.

The presentation of MHC class I-bound and MHC class II-bound peptides is important for initial immunization to an antigen but also for re-stimulating memory T cells formed during previous antigen exposure. We compared the anti-viral memory CD8⁺ T cell response in ACE 10/10 and WT mice that were challenged with a vaccinia virus strain modified to express a major polyoma MHC class I epitope 1 month after an initial polyoma infection. This approach tests the ability of anti-polyoma memory T cells to recognize and kill cells infected with the modified vaccinia virus. Four days after vaccinia virus infection, viral particles could not be detected in six of eight ACE 10/10 mice compared with only one of eight WT mice⁶⁴. By contrast, an equivalent experiment performed with unmodi fied vaccinia virus (a strain that does not express the polyoma epitope) showed no difference between the two groups. These experiments suggest that ACE over-expression potentiates the anti-viral memory CD8⁺ T cell response. Mechanistically, we do not yet know whether this potentiation is due to changes in antigen presentation or alterations in how APCs interact with T cells. Understanding the precise mechanism may contribute to novel immunization strategies. Antigen processing also has a role in antibody production, as the humoral immune response depends on antigen presentation by APCs, B cells that recognize the antigen and CD4⁺ T helper cells (either T_H1 or T_H2) that help B cells mature and proliferate. To test whether ACE expression levels in macrophages affect the humoral immune response, we used an enzyme-linked immunosorbent assay (ELISA) to measure antibody concentrations in plasma in ACE-deficient, WT and ACE 10/10 mice immunized with OVA⁶⁵. We also compared the ability of these macrophages to stimulate CD4⁺ T cells. ACE 10/10 mice consistently produced more anti-OVA antibody than WT mice; for example, IgG1 levels in ACE 10/10 mice were over 20-fold higher than those in WT mice. Furthermore, ACE expression levels in macrophages positively correlated with their ability to stimulate OVA-sensitive CD4⁺ T cells ex vivo. ACE overexpression in these cells therefore facilitates the processing and display of OVA peptides bound to MHC class II proteins, which stimulate CD4⁺ T cells (FIG. 3b).

The extent to which ACE trimming of both MHC class I and MHC class II peptides affects antigen presentation depends on the amino acid sequence of the individual peptides. For example, increased ACE expression increases MHC class I OVA epitope (SIINFEKL) presentation but decreases presentation of the MHC class I polyoma virus middle T antigen (RRLGRTLLL)⁶². Similarly, ACE overexpression differentially affected the display of four hen egg lysozyme MHC class II epitopes: peptide presentation was increased for one peptide, decreased for two peptides and had no effect on the fourth peptide⁶⁵. Given these data, one cannot conclude that ACE always increases the immunogenicity of a particular peptide. Rather, it seems likely that as cells change their production of ACE, the peptides they display will also change. This may ultimately increase the immune response through the simple mechanism of displaying different epitopes and engaging a different subset of CD4⁺ T_H1 or T_H2 cell clones.

ACE and neutrophil function.

Neutrophils are among the first cells to respond to a bacterial infection, and they use a variety of mechanisms — such as phagocytosis, ROS generation, antibacterial peptides and neutrophil extracellular trap (NET) formation (see below) — to destroy the invading bacteria^66,67. WT neutrophils increase ACE expression levels in response to an MRSA-mediated immune challenge⁶⁸. To study the effect of neutrophil ACE overexpression in response to bacterial infection, we generated a transgenic mouse model (NeuACE mice) in which so-called NeuACE neutrophils overexpress ACE by 12-fold to 18-fold compared with cells from WT mice⁶⁸ (FIG. 4a). NeuACE mice have substantially increased resistance to infections of MRSA, Klebsiella pneumoniae and Pseudomonas aeruginosa. For example, when challenged with MRSA infection, the lesion size and bacterial counts in NeuACE mice were only ~25% of those observed in WT mice (FIG. 4b–d). This phenotype is dependent on neutrophil function, as neutrophil elimination with anti-neutrophil antisera eliminates the differences in both lesion size and bacterial counts between NeuACE and WT animals⁶⁸. In addition, NeuACE neutrophils kill bacteria more efficiently than do WT neutrophils in vitro. By contrast, ACE-deficient neutrophils are significantly less effective at killing bacteria than are WT cells in vitro and in vivo: for example, at the end of an in vitro whole-blood killing assay, bacterial counts in samples with ACE-deficient neutrophils were 6-fold greater than in those using WT neutrophils and 30-fold greater than in those using NeuACE neutrophils⁶⁸. These data suggest that a direct relationship between ACE production in neutrophils and the ability to kill bacteria exists.

Open in a separate window

Figure 4 |

ACE overexpression in neutrophils reduces skin lesions caused by MRSA infection.

a | NeuACE mice overexpress angiotensin-converting enzyme (ACE) in neutrophils, but the enzyme is present in other cell types as well. b | Skin lesion area sizes (and standard error of the mean) of NeuACE and wild-type (WT) mice that were subcutaneously injected with methicillin-resistant Staphylococcus aureus (MRSA) on day 0, including mice that had previously been depleted of neutrophils using anti-neutrophil antibodies. The data show that overexpression of ACE in neutrophils provides resistance against MRSA infection. c | Bacterial numbers within the skin lesions after 3 days in NeuACE and WT mice show that overexpression of ACE in neutrophils attenuates bacterial numbers following MRSA infection. Data from individual mice (open blue diamonds) as well as the group means (orange bars) and standard error of the mean are shown. d | Representative images of NeuACE and WT mice 3 days after infection with MRSA, illustrating the difference in lesion size. *P < 0.001. CFU, colony forming unit; NS, not significant. Parts b–d are republished with permission of American Society of Hematology, from Khan, Z. et al. Angiotensin-converting enzyme enhances the oxidative response and bactericidal activity of neutrophils. Blood 130, 328–339 (2017); permission conveyed through Copyright Clearance Center, Inc. (REF. 68).

ACE and superoxide production.

One of the central ways by which neutrophils destroy bacteria is through phagocytosis, followed by fusion of the phagocyte with cytoplasmic granules. The enclosed bacteria are killed by several different mechanisms, but one of the most powerful is the generation of superoxide (O₂⁻) by NADPH oxidase. O₂⁻ is then converted to hydrogen peroxide (H₂O₂) and hypochlorous acid (HClO), which are bactericidal⁶⁹. To evaluate whether ACE has a role in O₂⁻ production, we challenged neutrophils from NeuACE and WT mice with MRSA and found that, although both cell populations had increased O₂⁻ production, levels from NeuACE cells were >2-fold higher than those from WT cells⁶⁸. By contrast, O₂⁻ production by ACE-deficient neutrophils was ~50% that of WT neutrophils⁶⁸. These changes correspond to the increased bacterial killing by NeuACE neutrophils and the decreased bacterial killing by ACE-deficient neutrophils described above. To verify that the increased bacterial killing in NeuACE neutrophils is due to increased O₂⁻, O₂⁻ production was blocked in NeuACE and WT neutrophils using the NADPH oxidase inhibitor gp91 ds-tat. This treatment resulted in bacterial counts that were comparable between the two cell types. O₂⁻ production in ACE 10/10 macrophages was also over twofold greater than in macrophages from WT mice and thus appears to be a common feature of ACE over-expression in myeloid cells (FIG. 5). Taken together, these data suggest that increased O₂⁻ production is a major mechanism that enables improved killing of bacteria by NeuACE neutrophils.

Open in a separate window

Figure 5 |

ACE overexpression enhances the adaptive and innate immune response.

a | In ACE 10/10 mice, angiotensin-converting enzyme (ACE) overexpression in monocytes and macrophages enhances the immune response in several ways. Resistance to bacterial infection is increased through elevated production of superoxide (O₂⁻), nitric oxide, tumour necrosis factor (TNF) and IL-12β; resistance to tumour growth is due to direct cytotoxic effects on tumour cells through an enhanced CD8⁺ T cell response and an increase in pro-inflammatory cytokines. In addition, ACE 10/10 mice produce more antibodies than do wild-type animals, presumably owing to an enhanced display of cell surface epitopes that increase the response of CD4⁺ T helper 1 (T_H1) or T_H2 cells, which have a key role in B cell maturation. b | In NeuACE mice, neutrophils overexpress ACE, leading to increased production of O₂⁻ in response to bacterial infection via phosphorylation of p47-PHOX, the regulatory subunit of the NADPH complex. Increased O₂⁻ production improves cytokine and neutrophil extracellular trap (NET) formation. IgG, immunoglobulin-γ; MHC, major histocompatibility complex; TCR, T cell receptor.

Neutrophils kill bacteria using mechanisms besides phagocytic killing, such as by releasing extracellular fibres termed NETs^70–72, which is stimulated by ROS generation⁷³. These fibres are composed of DNA and proteins that coalesce and act to entrap and consequently kill bacteria. NET release by NeuACE neutrophils (as measured by extracellular release of neutrophil elastase) is increased by tenfold in vitro compared with WT neutrophils⁶⁸. The increase was reverted to WT levels with inhibitors of either ROS generation or ACE activity. These data show that the ACE-mediated increase in ROS production improves bacterial killing by several mechanisms.

As indicated, neutrophil O₂⁻ is produced by NADPH oxidase, a membrane-bound enzyme composed of several protein chains. This enzyme is activated when one of the protein chains called p47-PHOX (also known as NCF1) — normally located in the cytoplasm — is phosphorylated and recruited to the cell or phagocytic membrane, where it assembles with other protein chains, forming the active enzyme. We have verified, using western blots, that in NeuACE neutrophils, the increased O₂⁻ production is due to increased phosphorylation of the p47-PHOX subunit. However, the exact biochemical link between increased cellular ACE activity and increased phosphorylation of p47-PHOX is still under active investigation. One approach focuses on identifying the kinase(s) activated by ACE that phosphorylates p47-PHOX. Another approach is to investigate which peptide(s) cleaved by ACE can affect NADPH oxidase activity. Finally, studies in NeuACE mice support the conclusion from ACE 10/10 mice that the observed phenotype is independent of angiotensin II, the inflammatory mediator bradykinin or the anti-inflammatory peptide acetyl-SDKP⁶⁸. Currently, the peptide(s) responsible for the increased immune effect in ACE 10/10 and NeuACE mice are not known.

The authors thank B. Taylor of Cedars-Sinai Medical Center for help in preparing the manuscript before submission. The authors’ work described in this Review was supported by US National Institute of Health grants P01HL129941, R21AI114965 and R03DK101592, and American Heart Association grants 16SDG30130015 and 17GRNT33661206.