A number of simian and simian human immunodeficiency viruses (SIV and SHIV, respectively) have been used to assess the efficacy of HIV-1 vaccine strategies. Among these, SIVmac239 is considered among the most stringent because, unlike SHIV models, its full genome has coevolved in its macaque host and its tier 3 envelope glycoprotein (Env) is exceptionally hard to neutralize. Here, we investigated the ability of eCD4-Ig, an antibody-like entry inhibitor that emulates the HIV-1 and SIV receptor and coreceptor, to prevent SIVmac239 infection. We show that rh-eCD4-IgI39N expressed by recombinant adeno-associated virus (AAV) vectors afforded four rhesus macaques complete protection from high-dose SIVmac239 challenges that infected all eight control macaques. However, rh-eCD4-IgI39N–expressing macaques eventually succumbed to serial escalating challenge doses that were 2, 8, 16, and 32 times the challenge doses that infected the control animals. Despite receiving greater challenge doses, these macaques had significantly lower peak and postpeak viral loads than the control group. Virus isolated from three of four macaques showed evidence of strong immune pressure from rh-eCD4-IgI39N, with mutations located in the CD4-binding site, which, in one case, exploited a point-mutation difference between rh-eCD4-IgI39N and rhesus CD4. Other escape pathways associated with clear fitness costs to the virus. Our data report effective protection of rhesus macaques from SIVmac239.

Adeno-associated virus (AAV) delivery of potent and broadly neutralizing antibodies (bNAbs is a promising approach for the prevention of HIV-1 infection. The immunoglobulin G (IgG)1 subtype is usually selected for this application, because it efficiently mediates antibody effector functions and has a somewhat longer half-life. However, the use of IgG1-Fc has been associated with the generation of anti-drug antibodies (ADAs) that correlate with loss of antibody expression. In contrast, we have shown that expression of the antibody-like molecule eCD4-Ig bearing a rhesus IgG2-Fc domain showed reduced immunogenicity and completely protected rhesus macaques from simian-HIV (SHIV)-AD8 challenges. To directly compare the performance of the IgG1-Fc and the IgG2-Fc domains in a prophylactic setting, we compared AAV1 expression of rhesus IgG1 and IgG2 forms of four anti-HIV bNAbs: 3BNC117, NIH45-46, 10-1074, and PGT121. Interestingly, IgG2-isotyped bNAbs elicited significantly lower ADA than their IgG1 counterparts. We also observed significant protection from two SHIV-AD8 challenges in macaques expressing IgG2-isotyped bNAbs, but not from those expressing IgG1. Our data suggest that monoclonal antibodies isotyped with IgG2-Fc domains are less immunogenic than their IgG1 counterparts, and they highlight ADAs as a key barrier to the use of AAV1-expressed bNAbs.

The human immunodeficiency virus type 1 (HIV-1) entry inhibitor eCD4-Ig is a fusion of CD4-Ig and a coreceptor-mimetic peptide. eCD4-Ig is markedly more potent than CD4-Ig, with neutralization efficiencies approaching those of HIV-1 broadly neutralizing antibodies (bNAbs). However, unlike bNAbs, eCD4-Ig neutralized all HIV-1, HIV-2, and simian immunodeficiency virus (SIV) isolates that it has been tested against, suggesting that it may be useful in clinical settings, where antibody escape is a concern. Here, we characterize three new eCD4-Ig variants, each with a different architecture and each utilizing D1.22, a stabilized form of CD4 domain 1. These variants were 10- to 20-fold more potent than our original eCD4-Ig variant, with a construct bearing four D1.22 domains (eD1.22-HL-Ig) exhibiting the greatest potency. However, this variant mediated less efficient antibody-dependent cell-mediated cytotoxicity (ADCC) activity than eCD4-Ig itself or several other eCD4-Ig variants, including the smallest variant (eD1.22-Ig). A variant with the same architecture as the original eCD4-Ig (eD1.22-D2-Ig) showed modestly higher thermal stability and best prevented the promotion of infection of CCR5-positive, CD4-negative cells. All three variants, and eCD4-Ig itself, mediated more efficient shedding of the HIV-1 envelope glycoprotein gp120 than did CD4-Ig. Finally, we show that only three D1.22 mutations contributed to the potency of eD1.22-D2-Ig and that introduction of these changes into eCD4-Ig resulted in a variant 9-fold more potent than eCD4-Ig and 2-fold more potent than eD1.22-D2-Ig. These studies will assist in developing eCD4-Ig variants with properties optimized for prophylaxis, therapy, and cure applications. HIV-1 bNAbs have properties different from those of antiretroviral compounds. Specifically, antibodies can enlist immune effector cells to eliminate infected cells, whereas antiretroviral compounds simply interfere with various steps in the viral life cycle. Unfortunately, HIV-1 is adept at evading antibody recognition, limiting the utility of antibodies as a treatment for HIV-1 infection or as part of an effort to eradicate latently infected cells. eCD4-Ig is an antibody-like entry inhibitor that closely mimics HIV-1's obligate receptors. eCD4-Ig appears to be qualitatively different from antibodies, since it neutralizes all HIV-1, HIV-2, and SIV isolates. Here, we characterize three new structurally distinct eCD4-Ig variants and show that each excels in a key property useful to prevent, treat, or cure an HIV-1 infection. For example, one variant neutralized HIV-1 most efficiently, while others best enlisted natural killer cells to eliminate infected cells. These observations will help generate eCD4-Ig variants optimized for different clinical applications.

Broadly neutralizing antibodies (bNAbs) can prevent and control an HIV-1 infection, but their breadth is invariably too limited for use as monotherapy. To address this problem, bi- and trispecific antibody-like constructs have been developed. These engineered antibodies typically have greater breadth than the native bNAbs from which they were derived, but they are not more potent because they do not, in most cases, simultaneously engage more than a single epitope of the HIV-1 envelope glycoprotein (Env). Here, we describe a new class of bispecific antibodies targeting the V2-glycan (apex) and V3-glycan regions of the HIV-1 envelope glycoprotein (Env). Specifically, bispecific antibodies with a single-chain (scFv) form of the CAP256.VRC26.25 V2-glycan (apex) antibody on one antibody arm and a full V3-glycan Fab on the other arm neutralizes more HIV-1 isolates than the bNAbs from which they were derived. Moreover, these bispecific antibodies are markedly more potent than their parental bNAbs, likely because they simultaneously engage both the apex and V3-glycan epitopes of Env. Our data show that simultaneous engagement of two critical epitopes of a single Env trimer can markedly increase the potency of a bispecific antibody. IMPORTANCE Broadly neutralizing antibodies (bNAbs) can prevent a new HIV-1 infection and can at least temporarily suppress an established infection. However, antibody-resistant viruses rapidly emerge in infected persons treated with any single bNAb. Several bispecific antibodies have been developed to increase the breadth of these antibodies, but typically only one arm of these bispecific constructs binds the HIV-1 envelope glycoprotein trimer (Env). Here, we develop and characterize bispecific constructs based on well-characterized V2-glycan and V3-glycan bNAbs and show that at least one member of this class is more potent than its parental antibodies, indicating that they can simultaneously bind both of these epitopes of a single Env trimer. These data show that bispecific antibody-like proteins can achieve greater neutralization potency than the bNAbs from which they were derived.

The engineered antibody-like entry inhibitor eCD4-Ig neutralizes every human immunodeficiency virus type 1 (HIV-1), HIV-2, and simian immunodeficiency virus isolate it has been tested against. The exceptional breadth of eCD4-Ig derives from its ability to closely and simultaneously emulate the HIV-1 receptor CD4 and coreceptors, either CCR5 or CXCR4. Here we investigated whether viral escape from eCD4-Ig is more difficult than that from CD4-Ig or the CD4-binding site antibody NIH45-46. We observed that a viral swarm selected with high concentrations of eCD4-Ig was increasingly resistant to but did not fully escape from eCD4-Ig. In contrast, viruses selected under the same conditions with CD4-Ig or NIH45-46 fully escaped from those inhibitors. eCD4-Ig-resistant viruses acquired unique changes in the V2 apex, V3, V4, and CD4-binding regions of the HIV-1 envelope glycoprotein (Env). Most of the alterations did not directly affect neutralization by eCD4-Ig or neutralizing antibodies. However, alteration of Q428 to an arginine or lysine resulted in markedly greater resistance to eCD4-Ig and CD4-Ig, with correspondingly dramatic losses in infectivity and greater sensitivity to a V3 antibody and to serum from an infected individual. Compensatory mutations in the V3 loop (N301D) and in the V2 apex (K171E) partially restored viral fitness without affecting serum or eCD4-Ig sensitivity. Collectively, these data suggest that multiple mutations will be necessary to fully escape eCD4-Ig without loss of viral fitness. IMPORTANCE HIV-1 broadly neutralizing antibodies (bNAbs) and engineered antibodylike inhibitors have been compared for their breadths, potencies, and in vivo half-lives. However, a key limitation in the use of antibodies to treat an established HIV-1 infection is the rapid emergence of fully resistant viruses. Entry inhibitors of similar breadths and potencies can differ in the ease with which viral escape variants arise. Here we show that HIV-1 escape from the potent and exceptionally broad entry inhibitor eCD4-Ig is more difficult than that from CD4-Ig or the bNAb NIH45-46. Indeed, full escape was not observed under conditions under which escape from CD4-Ig or NIH45-46 was readily detected. Moreover, viruses that were partially resistant to eCD4-Ig were markedly less infective and more sensitive to antibodies in the serum of an infected person. These data suggest that eCD4-Ig will be more difficult to escape and that even partial escape will likely extract a high fitness cost.