ACE2 immunoadhesin strategy
A potentially more promising strategy would be to create an antibody-like molecule that would bind to the coronavirus itself, rather than shielding cells from being infected. For this strategy, it is proposed to use a soluble version of the ACE2 receptor that would bind to the S protein of 2019-nCoV thereby neutralizing the virus ( Figure 1). Again, the research on the SARS virus suggests this strategy is potentially promising. Soluble ACE2 receptor was demonstrated to block the SARS virus from infecting cells in culture 42. The reported affinity of soluble ACE2 for the SARS S protein was 1.70 nM, which is comparable to the affinities of monoclonal antibodies 55; it is likely that 2019-nCoV has similar affinity for ACE2. In order to use ACE2 as a therapy to treat patients, it would be advisable to convert soluble ACE2 into an immunoadhesin format fused to an immunoglobulin Fc domain (ACE2-Fc), thereby extending the lifespan of the circulating molecule, while also recruiting effector functions of the immune system against the virus. While not tested in an animal model, a previous study demonstrated that an ACE2 extracellular domain fused to the human IgG1 domain (ACE2-NN-Ig) was effective in neutralizing SARS coronavirus in vitro, with a 50% inhibitory concentration of 2 nM 56. This study provides evidence then that ACE2-Fc could similarly inhibit 2019-nCoV in vitro and potentially in patients.
An additional advantage of using ACE2 as a 2019-nCoV S protein neutralizing agent is that ACE2 administration could also directly treat the pneumonia pathophysiology. A portion of patients with SARS and 2019-nCoV infection develop pneumonia, which is characterized by pulmonary edema and acute respiratory distress syndrome (ARDS) 1, 2. The viruses may, in part, cause ARDS through viral-induced ACE2 protein shedding and ACE2 protein decreased expression, both of which are mediated by S protein binding 54. Administration of recombinant ACE2 protein has been shown to improve acute lung injury through decreasing angiotensin II levels and the hormones subsequent binding to angiotensin II type 1a receptor 57. Recombinant ACE2 can also reduce ARDS in respiratory syncytial virus 58 and H5N1 influenza 59 infection models. Based on these promising preclinical studies, recombinant human ACE2 (rhACE2) was moved into clinical trials in order to treat ARDS in critically ill patients. A phase I trial demonstrated rhACE2 was well tolerated with no effects seen on the cardiovascular system 60. A phase II trial demonstrated on-target efficacy in reducing Ang1-8 peptide levels, but did not show significant modulation of respiratory parameters 61. It remains to be seen whether rhACE2 administration has the same clinical benefits in treating ARDS that have been seen in animal models, and whether ACE2-Fc administration could alleviate ARDS in 2019-nCoV patients.
The proposed therapy for 2019-nCoV patients would consist of the extracellular domain of the ACE2 protein fused to a human immunoglobulin G Fc domain ( Figure 2A). Studies have shown that the ACE2 amino acids 18 – 615 appear to be sufficient for SARS S protein binding 62, which also covers the peptidase domain necessary for ACE2 enzymatic function. It is possible a smaller portion of the extracellular ACE2 domain would be adequate for S protein binding, although a smaller version would lack enzyme activity beneficial in treating lung injury. Further studies are needed to define the minimal ACE2 domain necessary for 2019-nCoV S protein binding to construct even smaller ACE2-Fc proteins. While we do not know the structure of the 2019-nCoV S protein or how it binds to the ACE2 receptor yet, it is reasonable for now to assume that the same ACE2 protein domains utilized by the SARS virus are also bound by 2019-nCoV to infect cells.
Figure 2.  TherapeuticDesign of the ACE2-Fc fusion protein as a therapy against 2019-nCoV coronavirus.
( A) The extracellular domain of ACE2 is appended onto the human immunoglobulin Fc domain, including the hinge region. The Fc domain facilitates dimerization of two ACE2 domains. ( B) The amino acid sequence of the ACE2-Fc fusion protein is provided. The ACE2 domain consists of amino acids 18–615 of the human ACE2 protein (blue; UniProtKB - Q9BYF1). The sequence of the human immunoglobulin G isotype 1 constant region is provided (green; UniProtKB - P01857). A secretion signal from a human immunoglobin heavy chain is provided (red; UniProtKB - A0A0C4DH39). The advantage of the Fc domain is endowing a longer-half life of the drug, which could enable healthcare workers to potentially be given drug doses prophylactically before seeing infected patients. Indeed, the half-life of recombinant ACE2 was extended from less than two hours to over one week in mice when formatted as a recombinant ACE2-Fc therapy in a study evaluating treatment for hypertension 63. One difference from the prior blocking agent strategies is that the effector functions of the Fc domain could be retained in this molecule, allowing recruitment of dendritic cells, macrophages, and natural killer cells through the CD16 receptor against viral particles or infected cells. This may facilitate faster activation of the host antiviral immune response and elimination of the virus, which was illustrated in a SARS mouse model where Fc engaging antibodies were more potent in eliminating SARS via activation of phagocytic cells compared to antibodies that neutralized virus alone 64. Overall, the ACE2-Fc fusion protein would have many of the same benefits of a traditional neutralizing antibody that would be sought as a treatment for the infection, but represent one with maximal breadth and potency since the 2019-nCoV could not escape its neutralization, given the same protein is also its receptor for cell entry. Indeed, it has been shown that the pathogenicity of SARS versus the more mild human coronavirus NL63 was related to a lower affinity of NL63 for human ACE2 versus SARS, with NL63 S protein reducing ACE2 levels less than SARS S protein 65. Therefore, if 2019-nCoV were to try to escape ACE2 neutralization via decreasing affinity, it would mutate into a less pathogenic virus. This is similar to the re-emergent SARS virus in 2003-2004, which had lower affinity for human ACE2 and resulted in less severe infection and no secondary transmission 66. Thus, 2019-nCoV could be presented with an evolutionary trap when faced with potential ACE2-Fc therapy, leading toward a more benign clinical course.
To give some additional support to the potential of a receptor-immunoadhesin being a potential antiviral strategy, it should be noted that CD4-Fc or CD4-IgG was one of the early agents developed as a potential HIV medication 67. The protein contained the first two domains of the CD4 receptor that are known to bind gp120 on the surface of infected HIV cells. CD4-IgG was shown to neutralize HIV in vitro, preventing infection. The protein was also safe when administered in patients, although only limited-to-mild clinical benefit was achieved 68, 69. Updated enhanced versions of CD4-IgG have been developed that additionally have a small peptide derived from the co-receptor, CCR5, enhancing affinity and giving even more potent neutralizing activity, essentially 100% of HIV isolates and making rhesus macaques resistant to multiple simian-human immunodeficiency virus challenges 70, 71. While HIV and 2019-nCoV are very different viruses, with different cell types, kinetics, and clinical courses, the previous results with HIV are encouraging that this could be a therapeutic strategy for 2019-nCoV. If anything, 2019-nCoV is likely more amenable to this neutralizing therapy given that the respiratory virus will only cause an acute infection, unlike HIV, which causes chronic infection in hosts with different cellular reservoirs.
One potential limitation of the ACE2-Fc strategy is that the increase in levels of extracellular ACE2 could have unknown effects on the body, particularly when elevated for a prolonged time via Fc domain extended half-life. Small levels of extracellular ACE2 are already secreted by tissues, so the circulation of this extracellular domain would not be unprecedented 72. Moreover, recombinant ACE2 protein was well-tolerated by healthy patients in a phase I trial, and by patients with lung injury in a phase II trial, suggesting treating 2019-nCoV patients with ACE2-Fc will also tolerated. If investigators are still concerned, critical amino acid(s) for ACE2 peptidase activity could be mutated to abolish the native function of this sequence, while retaining high affinity binding for SARS and 2019-nCoV S protein. Indeed, this possibility was previously investigated in generating an ACE2 and IgG1 fusion protein, which showed that mutation of histidine residues at position 374 and 378 of the ACE2 extracellular domain abolished peptidase activity, but retained high affinity binding to SARS S protein 56. Of course, ACE2 peptidase mutation would eliminate the beneficial effects from the recombinant protein delivery of ACE2 in treating lung injury, so it is recommended that retaining ACE2 enzyme activity be pursued first. Another potential concern is that receptor binding via an antibody format could inadvertently direct 2019-nCoV toward infecting Fc receptor (CD16) positive cells, which has been shown in vitro for neutralizing antibodies in MERS 73. It’s unclear what clinical significance this would have, and to what extent this would happen in vivo. Ultimately, clinical trials will be needed to delineate any specific side effects of ACE2-Fc treatment.