Angiotensin-converting enzyme 2 (ACE2) was discovered in 2000 as a homologue of ACE, insensitive to ACE inhibitors 1, 2. Like its homologue ACE, ACE2 is anchored in the cell membrane with its enzymatic domain outside of the cell as what is termed an “ectoenzyme” hence ideally oriented for cleaving certain regulatory peptides in extracellular space. Both ACE2 and ACE are also cleaved (shed) from the plasma membrane. ACE2 is known to be cleaved by a disintegrin and metalloprotease 17, called ADAM17, whereas the “sheddase” that cleaves ACE is unknown. Donoghue et al 2 found that ACE2, like ACE, targets angiotensin (Ang) I but instead of generating Ang II from Ang I, like ACE does, ACE2 converts Ang I to Ang-(1–9). The same year, another group of investigators, led by Anthony Turner, reported that ACE2 cleaves not only Ang I to Ang-(1–9) but also cleaves Ang II to form Ang-(1–7) 1. This was an important discovery because in doing so ACE2 prevents accumulation of Ang II and attenuates the undesirable effects of excess of this octapeptide such as fibrosis, inflammation and oxidative stress while fostering the counter-regulatory actions via formation of Ang-(1–7) 3–6. Since the discovery of ACE2 there has been an increased interest in the so called ACE2-Ang-(1–7) axis as an alternative pathway to the long-recognized ACE-Ang II axis 3–11.Early on our lab focused on the localization of ACE2 within the kidney demonstrating ACE2 expression in podocytes and the proximal tubular brush border and postulated that ACE2 could have renoprotective properties 12, 13. Such protection is conferred largely at the tissue level because under normal conditions, the circulating levels of ACE2 are very low 14. Of note, the circulating level of ACE2 in plasma may increase significantly in cardiovascular disease and has potential as a biomarker 15. ACE2 is a “tissue” enzyme present in the cell membrane of many organs, most abundantly in epithelial cells of the small intestine and the kidney proximal tubule 1, 2, 12, 13, 16. In the lungs, ACE2 is expressed on type II pneumocytes that represent an important entry site for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) 17, 18. ACE2 is one of three enzymes known to convert Ang II to Ang-(1–7). The other two are Prolylcarboxypeptidase (PRCP) and Prolyloligopeptidase (POP) 14, 19. Of these three enzymes, ACE2 has the highest efficiency to catalyze the conversion of Ang II to Ang-(1–7) 14, 20. PRCP is also expressed highly in the kidney but in the collecting tubule, predominantly at the apical site of α-intercalated cells, rather than in the proximal tubule where ACE2 is expressed 19. In the circulation and the lungs, the prevailing enzyme that degrades Ang II into Ang-(1–7) is POP 14. POP can also directly form Ang-(1–7) from Ang I but neprilysin appears to have higher efficiency for this reaction 21, 22.This review honors Dr. Donald Seldin’s groundbreaking contributions to renal physiology and kidney disease, and it is based on the Donald Seldin Lecture Award, of the AHA Hypertension Council, given to one of us (DB) in Boston, September 2023. Dr. Seldin is one of the fathers of nephrology, and a giant in medicine of the past century. He elevated the pathophysiology of kidney disease to an art and his teachings and contributions are still the foundation for what we teach our medical students and fellows in current textbooks. He transformed the department of medicine at Southwestern University, Dallas, and trained a vast number of nephrology fellows that later became prominent nephrologists and chiefs of nephrology departments in the US and abroad. The review discusses published work, mainly from our laboratory on ACE2 focusing on the following aspects: 1) Distinction between full-length membrane-bound ACE2 and different soluble ACE2 proteins, 2) Therapeutic potential of soluble ACE2 proteins for kidney disease, and 3) The critical role of ACE2 as the essential receptor for SARS-CoV-2 and the therapeutic/preventative opportunities of soluble ACE2 proteins to combat infections caused by SARS-CoV-2.Membrane-bound full-length ACE2 consists of 805 amino acids, where amino acids 18–615 comprise the peptidase domain and amino acids 616–768 represent the collectrin-like domain that mediates dimerization 1, 2, 23–25. Amino acids 740–768 represent the transmembrane domain and amino acids 769–805 represent a C-terminal cytoplasmic tail 1, 2, 23–25 (figure 1). The sequence of ACE2 consisting of amino acids 1–18, represents the signal peptide that directs the ACE2 to the plasma membrane 1, 2. Full-length ACE2 in the cell membrane exerts the catalytic action on peptides of the renin-angiotensin system (RAS) and other biological systems via its large extracellularly protruding 740 amino acid long ectodomain 10. Full-length ACE2, moreover has also been reported to be involved in gastrointestinal amino acid transport and cell signaling 26–29. In the kidney, however, a homologue of ACE2, collectrin, is responsible for amino acid reuptake 28, 30.Evidence pointing to a natural soluble ACE2 protein of 706–708 amino acids is largely indirect and based on the known cutting sites of proteases with the potential to cleave ACE2, as well as evidence gathered from in vitro studies 31–36. ADAM17, also called TACE (tumor necrosis factor-α-converting enzyme) can shed ACE2 from the plasma membrane of cells 33–35, 37, 38, the putative size of such cleaved soluble ACE2 would be ~ 708 amino acids 33, 36. Recently, ADAM10 has also been implicated as ACE2 sheddase, but the size of the formed soluble ACE2 by this protease in biological systems has not yet been shown 32. A recent report had suggested Membrane-type 1 Matrix Metalloproteinase (MT1-MMP) as the main sheddase of ACE2, cleaving ACE2 at the 706 amino acid residue 31. There is a marked overlap between MT1-MMP and ADAM10/17 in shedding properties for several substrates (meprin beta, interleukin-6 receptor and the amyloid precursor protein) making it likely that the catalytic redundancy of these enzymes could also include ACE2 as a substrate 32, 39.In addition to determining the exact size of the main naturally occurring form(s) of soluble ACE2, it is crucial to know whether it can form dimers or not since this has important implications for SARS-CoV-2 spike protein binding. As noted above soluble ACE2 740 is a dimer and it has been shown that binding to the SARS-CoV-2 spike protein by artificially dimerized soluble ACE2 proteins constructed as ACE2-Fc-fusion proteins is much higher than binding of their monomeric counterparts 40. Disruption of dimerization of cell membrane bound ACE2 receptor, moreover, markedly blocked SARS-CoV-2 pseudovirus binding to the host cell surface 41. The ACE2 peptide sequence from amino acid residues 616–658 was identified to be necessary for blocking ACE2 dimerization and the interaction between ACE2 and viral spike protein 41. However, this finding has not yet been confirmed independently by other groups. Altogether, there is strong evidence pointing to increased binding to SARS-CoV-2 and its neutralization when ACE2 is in dimeric form or at higher levels of multimerization 40, 42–45. This increased binding of such dimerized or multimerized ACE2 constructs may be primarily related not to increased affinity (a measure of the binding strength at a single binding site), but rather their increased avidity (a measure of the total binding strength) 40.The longest soluble ACE2 protein that was mainly generated by recombinant expression consists of the N-terminal 740 amino acids (referred here as ACE2 740). ACE2 740 lacks the transmembrane domain present in the full length membrane bound ACE2 805 25. The soluble ACE2 740 and the full length membrane bound ACE2 are both dimeric, because both contain a large portion of the dimerization prone collectrin-like domain 25, 46. Based on structural analysis, it has been assumed that the soluble ACE2 740 is the largest circulating form of ACE2. The actual size of the main circulating form of soluble ACE2 has not yet been fully delineated and there may be, in addition to ACE2 740, different soluble ACE2 proteins in the circulation 31, 32, 37, 47.The naturally occurring forms of soluble ACE2 can be found in cerebrospinal fluid, urine and in small amounts in the blood 14, 47–49 (figure 1). ACE2 enzymatic activity in urine is higher than in plasma and moreover is increased in certain forms of kidney disease 14, 50–54. For instance, ACE2 enzymatic activity levels about 10 times higher in urine than plasma were reported by us in two different mouse models of diabetes 51. In urine samples from patients with type 1 diabetes, urinary ACE2 is also increased 53. In a model of acute kidney injury (AKI) caused by unilateral ischemia reperfusion injury (IRI), there is an increase of urinary ACE2 activity that can be attributed to the presence of membrane bound ACE2 (molecular weight of about 100kDa) which originates from the damaged proximal tubules 54.Our lab found kidney ACE2 expression in podocytes and very strong expression in the brush border of proximal tubules 12, 13. Tubular ACE2 was increased in young diabetic db/db mice before the development of signs of diabetic kidney injury12. Glomerular expression of ACE2, by contrast, was reduced in diabetic (db/db) mice as compared to controls 13. Urinary albumin excretion increased in diabetic mice treated with a specific ACE2 inhibitor (MLN-4760) for 16 weeks as compared to controls 13. Concomitant administration of the AT1 blocker, Telmisartan, prevented the increase in urinary albumin excretion seen in mice that received only MLN-4760 13. This suggested that the effect of ACE2 was mediated by Ang II probably via the AT1 receptor. Further studies by our lab in the Streptozotocin (STZ) model of diabetes showed that inhibition of ACE2 enzymatic activity by MLN-4760 resulted in marked worsening of glomerular injury and overexpression of glomerular ACE 55. Based on the aggregate of these studies, we postulated that ACE2 exerts a renoprotective action by metabolizing Ang II and therefore attenuating the undesirable actions of excess of this peptide on the kidneys 9.Amplification of ACE2 enzymatic activity has been approached using several methods including ACE2 gene delivery (lentivirus, adenovirus and adeno-associated virus) 56–58, minicircle DNA delivery 52, transgenic rodent models 59, ACE2 activators 60, 61 and the administration of recombinant soluble ACE2 proteins (reviewed in 62). Different models of disease have been targeted experimentally such as Ang II mediated hypertension and myocardial fibrosis, pulmonary hypertension and diabetic kidney disease 52, 56, 57, 59, 63, 64. The administration of a soluble ACE2 protein (ACE2 740) was used first by the Penninger group in a piglet model of acute respiratory distress syndrome 65. This group also tested recombinant human soluble ACE2 740 in healthy volunteers and patients 66, 67, including patients with coronavirus disease 2019 (COVID-19) (see below) 68, 69. Here, we will discuss our work with recombinant soluble ACE2 proteins and the rationale for administration of shorter soluble ACE2 proteins in kidney disease.Administration of human soluble recombinant ACE2 with 740 amino acids (hACE2 740) to mice led to formation of neutralizing antibodies after ~2 weeks of administration 6, 70. To circumvent this issue we developed a murine form of recombinant soluble ACE2, also with 740 amino acids (mACE2 740), and tested it in a model of STZ induced diabetic kidney disease 52. In this model, daily intraperitoneal administration of mACE2 740 for several weeks, or delivery via minicircle DNA that resulted in very high levels of plasma ACE2 activity sustainable for months, did not improve albuminuria or kidney pathology 52. This lack of effect of mACE2 740 was noted despite a more than 100-fold increase in serum ACE2 enzymatic activity which, however, did not result in any increase in urine ACE2 enzymatic activity 52. This finding demonstrated that this form of soluble ACE2 cannot pass the glomerular filtration barrier; suggesting that kidney delivery of ACE2 to the intrarenal RAS is needed for its full therapeutic effect. This form of soluble ACE2, had no effect on blood pressure likely because in the STZ-model the levels of plasma Ang II are not increased 52. There was no evidence of toxicity in terms of body weight and overall health of mice with the long-term administration of soluble ACE2 (unpublished data).To accomplish delivery of soluble ACE2 to the kidney, our group bioengineered shorter soluble ACE2 proteins small enough to pass the glomerular filtration barrier while retaining full ACE2 enzymatic activity 71. These experiments were first done with mouse recombinant ACE2, that was truncated from the C-terminus to generate shorter ACE2 truncates with 619 amino acids (mACE2 619) and 605 amino acids (mACE2 605) 71. These recombinant shorter soluble ACE2 proteins retain full enzymatic activity, whereas the soluble ACE2 molecule shortened to 522 amino acids had no enzymatic activity 71. To demonstrate that mACE2 605 and mACE2 619 pass the glomerular filtration barrier, they were infused to ACE2-deficient mice 71. An increase in urine ACE2 enzymatic activity was detected in mice that received mACE2 605 or 619, but not in mice that received mACE2 740 71 (figure 2). Infusion with L-Lysine, a blocker of proximal tubular protein reabsorption, further increased urinary ACE2 activity in mice that received mACE2 605 or 619 demonstrating glomerular filtration and tubular uptake of mACE2 605 and 619 but not mACE2 740 (figure 2) 71. In addition, microSPECT/CT radioimaging of mice infused with recombinant soluble technetium-labeled mACE2 605 and 619 showed a clear uptake on nephrogram mainly in the cortical area of the kidney. By contrast, after infusion of the same dose of mACE2 740 there was no kidney uptake (Wysocki, Haney, and Batlle, unpublished).We next bioengineered a human recombinant soluble ACE2 protein with 618 amino acids, that we termed hACE2 618, as the human counterpart to the mACE2 619 protein 72. To increase the duration of action from hours to days, hACE2 618 was additionally fused with an albumin binding domain (ABD), and the resulting protein was termed hACE2 618-ABD 72. When hACE2 618-ABD was infused to wildtype mice there was an increase in urinary ACE2 enzymatic activity that was further enhanced after administration of L-Lysine, a blocker of proximal tubular protein absorption, demonstrating glomerular filtration and tubular reabsorption of hACE2 618-ABD (Wysocki et al, unpublished).As noted above, a rationale for the administration of recombinant soluble ACE2 proteins in kidney disease is based on the known role of ACE2 in metabolizing excess Ang II locally within the kidney 52, 62, 71, 73. Moreover, recent findings of kidney ACE2 deficiency in certain forms of AKI 54, 74 discussed below further support this potential therapeutic role. Activation of the RAS in the pathophysiology of AKI is supported by rodent and human studies showing increased urinary and/or kidney angiotensinogen, renin and Ang II levels 75–81. In a model of genetic ACE2 deficiency, moreover, inflammation, oxidative stress and apoptosis were exacerbated in AKI caused by IRI 81. Da Silveira et al also studied IRI induced AKI and found that at 4 hours post IRI, kidney Ang II levels were increased, whereas Ang-(1–7) and ACE2 mRNA were decreased 80.A strong rationale for administering soluble ACE2 comes from recent observations of our lab in a model of AKI with severe proximal tubular injury caused by unilateral IRI. In this model we found a striking maldistribution of ACE2 in kidneys harvested 48 hours post-injury (figure 3) 54. In addition, there was reduced ACE2 protein and enzymatic activity in cortical membranes of kidneys, with spillage of ACE2 into the tubular lumen and presence of ACE2-positive luminal casts in the medulla (figure 3) 54. Urinary ACE2 consistent with the size of the full-length membrane-bound ACE2, moreover, was increased 54. We also have found a decrease in apical proximal tubule ACE2 in a model of AKI caused by delayed graft function post-syngeneic kidney transplantation in mice 82.Nath et al studied ACE2 mRNA levels in AKI caused either by bilateral IRI or LPS-induced AKI and compared it to the model of reduced GFR caused by unilateral urinary tract obstruction (UUO) 74. The latter model was used to ascertain whether reduced GFR per se can influence kidney ACE2 mRNA expression. Kidney ACE2 mRNA was decreased in mice with ischemic AKI caused by bilateral IRI characterized by decreased GFR and abundant cell death ACE2 mRNA was unchanged in mice with reduced GFR caused by UUO 74. In our studies with IRI and post-kidney transplantation with delayed graft function, the main finding is a loss of ACE2 protein from the proximal tubule as a result of proximal tubule injury 54, 82.These observations support the concept of administrating recombinant shorter soluble ACE2 truncates to target the kidney RAS and counteract loss of kidney membrane-bound ACE2 in AKI. ACE2 deficiency favors the accumulation of Ang II with its pro-inflammatory and pro-fibrotic effects 52, 54, 62, 71, 83. Therefore, administration of a filterable form of soluble ACE2 offers a logical approach to prevent kidney Ang II accumulation in some forms of kidney disease. The resulting increase in Ang-(1–7) formation expected from ACE2 administration, moreover, may also exert an additive renoprotective effect. In our opinion, however, the main benefit comes from reducing Ang II levels whose actions are more powerful than those of Ang-(1–7).In January of 2020, Wan et al reported, based on structural comparisons of SARS-CoV and SARS-CoV-2 sequences, that ACE2 is the cell entry receptor for SARS-CoV-2 84, as it had been previously recognized for SARS-CoV 85. Before the SARS-CoV-2 outbreak was declared a pandemic, we hypothesized that soluble ACE2 proteins may act as a decoy by binding to SARS-CoV-2 and thus intercept the binding to membrane-bound ACE2, preventing cell entry of SARS-CoV-2 and subsequent viral replication 86. This decoy effect of soluble ACE2 for SARS-CoV-2 was demonstrated in permissive kidney cells and organoids by Monteil et al87, and shortly after by our group also using kidney organoids 72. Monteil et al showed that native human soluble ACE2 740 can inhibit SARS-CoV-2 infection of human vascular and kidney organoids in a dose-dependent manner, as shown by decreased SARS-CoV-2 RNA, measured by qRT-PCR 87. Wysocki et al showed that in human kidney organoids with proximal tubule-like structures there was a co-localization of ACE2 and transmembrane serine protease 2 (TMPRSS2) 72. Neutralization of SARS-CoV-2 infection in kidney organoids was then tested for native human and mouse soluble ACE2 740 and two shorter soluble ACE2 truncates, called ACE2 618 and ACE2 618-ABD designed in our laboratory 72. All three human soluble ACE2 proteins neutralized SARS-CoV-2 infection in human kidney organoids.Since SARS-CoV-2 variants do not bind mouse ACE2, in vivo studies in mice relied on transgenic mice, mainly the k18hACE2 transgenic model, in which human ACE2 expression is driven by the k18 promoter 88. When inoculated intranasally with a high dose of SARS-CoV-2, k18hACE2 mice develop severe disease with severe weight loss and lethality rates up to 100% by days 5–7 6, 42, 89–93. We have used this mouse model to examine the efficacy of a novel soluble ACE2 protein with enhanced binding affinity for SARS-CoV-2 42, 92, 94. To increase binding of our monomeric ACE2 618-ABD protein we used a dodecapeptide (DDC) motif which resulted in dimerization via disulfide bonds. The resulting protein was termed ACE2 618-DDC-ABD and its binding affinity for the SARS-CoV-2 receptor binding domain (RBD) was ~ 30-fold higher compared to its monomeric counterpart ACE2 618-ABD, and also improved as compared to the dimeric soluble ACE2 740 42.In an in vivo proof of concept study, ACE2 618-DDC-ABD was administered concomitantly via intranasal and intraperitoneal routes 1 hour before viral inoculation with 2×104 plaque forming units (PFU) wildtype SARS-CoV-2, followed by additional doses at 24 and 48 hours post-viral inoculation. Infected control mice received PBS at the same time points 42. Survival by day 14 was near complete in the group that received ACE2–618-DDC-ABD, whereas none of the mice from the control group survived past day 7 42.In a later study, we then compared intranasal to systemic administration of ACE2 618-DDC-ABD and also the efficacy of pre vs only post-inoculation administration 92. Intranasal administration 1 hour before and 24 and 48 hours post-viral inoculation with 2×104 PFU wildtype SARS-CoV-2 was superior to systemic (intraperitoneal) administration 92 (figure 4). Survival on day 5 was 90% in the intranasal group compared to 40% in the intraperitoneal group (figure 4, middle panel) 92. Administration of ACE2 618-DDC-ABD only after viral inoculation (at 24, 48 and 72 hours post) only provided partial protection from lethality 92 (figure 4, lower panel).Brain titers were undetectable in mice that received ACE2 618-DDC-ABD that survived until day 14; therefore, protection from brain SARS-CoV-2 invasion is a key determinant for survival in k18hACE2 mice 92. Of interest, the high mortality in untreated infected mice is seen despite only subtle brain histopathology. We found perivascular and leptomeningeal lymphocytosis, endothelial hyperthrophy and parenchymal inflammation, located mainly in the striatum, cerebral cortex and hypothalamus 92. Brain immunofluorescence staining for ionized calcium-binding adapter molecule 1 (IBA1) and glial fibrillary acidic protein (GFAP), markers for microglia and astrocytes, revealed high IBA1 and GFAP expression suggesting an underlying state of neuroinflammation (figure 5) 92. In great contrast, in k18hACE2 mice that received intranasal ACE2 618-DDC-ABD both before and after viral inoculation, brain histopathology appeared normal and IBA1 and GFAP immunofluorescence staining was markedly decreased (figure 5) 92. Protection from brain SARS-CoV-2 invasion and injury, therefore, was best achieved by intranasal administration of ACE2 618-DDC-ABD 92.In a subsequent study, we extended the time span from 1 hour to 6 hours between administration of ACE2 618-DDC-ABD and viral inoculation of k18hACE2 mice with SARS-CoV-2, with additional dosing at 24 and 48 hours post viral inoculation 94. These experiments were done with the aggressive delta SARS-CoV-2 variant 94. Similar to when inoculated with wildtype SARS-CoV-2, survival in the delta SARS-CoV-2 infected untreated control group was 0% by day 7, whereas survival in mice that received ACE2 618-DDC-ABD was near complete 94.Favorable attributes of an ACE2 decoy include high affinity for SARS-CoV-2, long duration of action, preserved ACE2 enzymatic activity and universal efficacy for all SARS-CoV-2 variants. Since the decoy hypothesis for SARS-CoV-2 was first proposed by us86 and the Penninger group95, different soluble ACE2 proteins have been bioengineered by various groups 42–45, 72, 87, 96–105. The discussion of all these different ACE2 proteins and their characteristics is beyond the scope of this review. Below we discuss the main variables regarding ACE2 protein structure, extending duration of action via fusion proteins and the presence or absence of ACE2 enzymatic activity.Some proposed ACE2 decoys consist of the complete, unaltered sequence of a native form of soluble ACE2 with 740 amino acids 43, 87, or were truncated at amino acid 618 or 615 resulting in monomers but have an unaltered amino acid sequence otherwise 42, 43, 72. It was found, however, that monomeric ACE2 proteins have low affinity for SARS-CoV-2 as compared to their dimeric counterparts 42–45. Other groups introduced mutations into the ACE2 amino acid sequence to increase the affinity for SARS-CoV-2 44, 45, 97, 98, 100–102. A different approach is the use of ACE2 tetramers to mirror the tetrameric structure of the S1 spike proteins 102, 103, or IgM-like penta- and hexameric soluble ACE2 proteins developed by Liu et al 104.The duration of action of native soluble ACE2 proteins, measured by plasma ACE2 activity, is only a few hours 42, 71, 72. To increase duration of action from hours to days, ACE2 proteins can be fused with an ABD 42, 72, 92, 94, Fc-tag 83, 97–99, 101, 102, 104, or domain I of human serum albumin (HSA) for FcRn binding 43. Most, if not all, therapeutic proteins, including Fc and HSA fusion proteins, are immunogenic and can induce anti-drug antibodies in humans 106. In addition, there has been some concern about the use of Fc-tags on soluble ACE2 proteins because of potential antibody dependent enhancement caused by Fc-receptor activation 107. Chen et al, however, found that administration of an ACE2-Fc decoy with active Fc-effector function was better than an ACE2-Fc decoy with inactive Fc-effector function in k18hACE2 mice 101. Although Fc fusion proteins have been shown to have a relatively low incidence of immunogenicity, there has been a report on an Fc fusion drug (rilonacept) with high anti-drug antibody titers 108. An anti-drug response was observed in 35% of patients, however, these responses were not associated with exaggerated clearance or loss of activity of the drug 108. Only a modest potential for neo-epitopes in HSA fusions currently marketed has been observed, which is mainly related to the linker fusion region and not epitopes of HSA itself 108. Our group used ABD for half-life extension strategy of soluble ACE2 in part because ABD when fused to an Affibody molecule did not increase immune response in long-term studies in rats or mice 109. In addition, the small size of the ABD-tag (~ 5 kDa) as opposed to Fc (50 kDa when dimerized) or HSA (65 kDa) only minimally increases the size of the protein to which the ABD tag is added. This is important to maintain the passage of the glomerular filtration barrier by short soluble ACE2 proteins as mentioned above for targeting of the kidney renin-angiotensin system.ACE2 enzymatic activity has been shown to be protective in experimental lung injury by dissipating of pro-inflammatory peptides like Ang II and des-Arg bradykinin 110, 111. It has been shown that with both SARS-CoV and SARS-CoV-2 infection there are decreased levels of membrane-bound full-length ACE2 because of internalization of the ACE2-SARS-CoV-2 complex 112–114. Administration of soluble ACE2 proteins in SARS-CoV-2 infection, therefore, replenishes ACE2 enzymatic activity and decreases the levels of unwanted pro-inflammatory peptides 95, 115. Some groups have proposed soluble ACE2 variants in which ACE2 enzymatic has been abrogated to use as decoys for treatment of SARS-Cov-2 infection 45, 97, 98, 101, 104, 105. The rationale for this choice was to avoid possible side effects related to ACE2 enzyme activity. This attribute, in our opinion, is a negative one as it is more likely that preserving ACE2 activity confers additional therapeutic benefit. Consistent with this view, Zhang et al compared administration by inhalation of a catalytically active vs inactive ACE2 decoy protein in k18hACE2 mice and found that the catalytically active soluble ACE2 proteins were superior in protecting from otherwise lethal SARS-CoV-2 infection 100.The above mentioned soluble ACE2 proteins have been tested either in vitro or in animal models of SARS-CoV-2 like the transgenic k18hACE2 mice, Syrian gold hamsters or non-human primates 42–45, 72, 87, 96–105. One protein, APN001, a soluble ACE2 protein with 740 amino acids and the same amino acid sequence as native soluble ACE2 740, has been tested in a clinical trial in which it was administered intravenously twice daily to patients hospitalized with COVID-19 68. The primary endpoint, a composite of all-cause death or invasive mechanical ventilation, however, was not significantly improved in those that received APN001 as compared to placebo 68. The results of a second clinical trial to assess safety, tolerability, pharmacokinetics and pharmacodynamics of nebulized APN001 are not yet available 69. As we demonstrated in k18hACE2 mice, best results are obtained when administering soluble ACE2 proteins intranasally and prior to infection 92.As early as January of 2020, ACE2 was reported as the main receptor for SARS-CoV-2 cell entry 84, 116–118. Cell entry of SARS-CoV-2 starts with binding of the RBD within the S1 spike protein to membrane bound ACE2, causing a conformational change that exposes the S2’ subunit within the S2 spike protein 119, 120. After binding of the S1 spike protein to ACE2, TMPRSS2, also present on the cell surface, is necessary for S2 protein cleavage, leading to membrane fusion and internalization of the ACE2-SARS-CoV-2 complex 118, 119. In the absence of TMPRSS2, Cathepsin B and L can also cleave the S2’ site in the endolysosome 118, 119, 121, 122. The accepted model of cell infectivity that was derived from these studies is still generally valid today.The essentiality of ACE2 as the cell entry receptor was demonstrated in ACE2 deficient kidney organoids. CRISPR-Cas9 genome editing was used by Garreta et al 123 to generate an ACE2 knockout human pluripotent stem cell line, resulting in human kidney organoids deficient of ACE2 as confirmed by immunofluorescence and western blot. Control wildtype and ACE2 knockout kidney organoids were then infected with SARS-CoV-2 123, 124. Immunofluorescence staining for the SARS-CoV-2 nucleopeptide and qPCR for SARS-CoV-2 mRNA was then performed to evaluate SARS-CoV-2 infectivity 123, 124. Immunofluorescence staining for the SARS-CoV-2 nucleopeptide was absent in the ACE2 knockout organoids whereas it was clearly detectable in wildtype organoids 123. SARS-CoV-2 mRNA levels were low in the ACE2 knockout organoids, and high in the wildtype organoids (figure 6) 124. The lack of infectivity in the ACE2 knockout models provides strong evidence for the essentiality of membrane-bound ACE2 as the cell entry receptor for SARS-CoV-2 123, 124. These findings were later confirmed by another group, that also generated ACE2 knockout kidney organoids and similarly found a lack of infectivity for SARS-CoV-2 in these organoids 125.The issue of infectivity or lack of by a spectrum of low concentrations of soluble ACE2 was also studied 124. Low concentrations of soluble ACE2 740 had no effect on SARS-CoV-2 infectivity, which was clearly abrogated by a very high concentration of ACE2 740 124. The finding, that exposure to very low concentrations of soluble ACE2 740 does not affect SARS-CoV-2 infectivity 124, is contrary to a previous publication claiming an increase in infectivity 126. Since soluble ACE2 lacks the transmembrane domain that is needed to anchor it to the cell membrane, after binding of SARS-CoV-2 to soluble ACE2, it is not likely to cause enhanced viral entry into the cell and therefore an increase in infectivity is not expected with any concentration of soluble ACE2 124.The main advantage of soluble ACE2 proteins for neutralization of SARS-CoV-2 is first their universality for current and future emerging SARS-CoV-2 variants 29, 127. This is because SARS-CoV-2 depends on its affinity to membrane bound ACE2 for cell entry and subsequent replication. Mutations that increase affinity for membrane bound ACE2 while simultaneously decreasing affinity for soluble ACE2 proteins are unlikely 96. Indeed, no mutational escape of SARS-CoV-2 has been found, after 15 passages of co-incubation of the virus with soluble ACE2, whereas mutational escape occurred after only 4 passages of co-incubation with a monoclonal antibody 97.Recently, one mutation in SARS-CoV-2, Y489H, was identified that causes escape from wildtype soluble ACE2, however, at the price of losing the ability to competitively replicate in host cells 98. This mutation was found to be extremely rare in SARS-CoV-2 sequence data bases, likely because of its minimal infectivity 98. Newly emerging mutated SARS-CoV-2 variants often are resistant to the existing monoclonal antibodies as for example for the omicron variant that emerged in the fall of 2021 and was shown to also escape from antibodies found in serum from patients with prior SARS-CoV-2 infection 128, 129.The provision of ACE2 enzymatic activity with administration of soluble ACE2 proteins may be a secondary advantage additive to the primary decoy effect 86, 115. This is because internalization of the ACE2-SARS-CoV-2 complex can lead to relative ACE2 deficiency 112–114. Exogenous supplementation of enzymatically active soluble ACE2 proteins, like ACE2 618-DDC-ABD, therefore, can help dissipating pro-inflammatory peptides like Ang II and des-Arg Bradykinin that accumulate in SARS-CoV-2 infection and foster lung injury 110, 111.Another key advantage of soluble ACE2 proteins for COVID-19 is that neutralization of SARS-CoV-2 can be achieved quickly as opposed to vaccines that require weeks for antibody formation 130. In the transgenic k18hACE2 mouse model we documented near complete protection from lethality when administered either 1 or 6 hours prior to viral inoculation 42, 92, 94.Despite gigantic improvements in the control of the COVID-19 pandemic, there are emerging outbreaks and SARS-CoV-2 infections are common and many go unreported 131–133. Recently, in summer 2024, cases and hospitalizations have been increasing 134. A major problem is the occurrence of long-COVID, a multisystemic condition experienced by patients after the acute phase of a SARS-CoV-2 infection, with symptoms that can affect a broad range of organ systems 135, 136. Several studies show that SARS-CoV-2 RNA or protein can be found throughout several organ systems including the intestine, central nervous system and lungs, several months or more than a year after acute SARS-CoV-2 infection 137–140. This may contribute and be associated with long-COVID 141. The risk for long-COVID was shown to be less in vaccinated persons as compared to unvaccinated persons 142. The cumulative incidence of long-COVID, moreover, was found to be lower during the omicron-era as compared to the delta or pre-delta era, but remains substantial, possibly due to the surge in infections during the omicron-era and lower numbers of vaccinations 142. As of September 2024, KP.3.1.1, a subvariant of the previously dominating KP.3 variant derived from the JN.1 omicron variant, is the most common circulating SARS-CoV-2 variant 143. KP.3 showed increased immune evasion in neutralization assays with convalescent serum and a higher effective reproduction number than JN.1 144. It is very likely that this variant as well as newly emerging variants can all be neutralized with appropriate soluble ACE2 proteins given the universality of their neutralizing effect on SARS-CoV-2. Animal models of SARS-CoV-2 infection other than k18hACE2 mice include hamsters, mouse adapted SARS-CoV-2 variants and a new mouse model with specifically humanized knock-in ACE2 receptors 145–148. These models that are non-lethal allow longer observation of SARS-CoV-2 infection that better resembles the human disease and study therapeutic interventions.The work revised in this paper highlights the importance of ACE2 for kidney disease, and the potential therapeutic role of ACE2 amplification for kidney disease. To date, most of the information on the potential therapeutic use of soluble ACE2 is based on animal experimentation. To our knowledge, the only developed clinical grade soluble ACE2 protein, which is a dimeric form with 740 amino acids, was tested in safety studies and small studies for lung disease and COVID-19. For the latter indication, the design of soluble ACE2 proteins with enhanced binding capacity for SARS-CoV-2 that can also be administered intranasally, such as our protein, ACE2 618-DDC-ABD, should provide an important addition to the armamentarium of therapies to prevent and treat SARS-CoV-2 and other emerging coronaviruses that use membrane-bound ACE2 as their cell entry receptor. Approval and commercialization of our soluble ACE2 proteins is a challenging process that is underway. It is noteworthy, nevertheless, that a recombinant ACE2 protein designed to downregulate the renin-angiotensin system within the kidney has been modified to combat the formidable health crisis provoked by COVID-19.
DB
Dr. Daniel Batlle
Nephrology · General NephrologyNorthwestern Medicinedanielb