Antivirals targeting paramyxovirus membrane fusion

Abstract

The Paramyxoviridae family includes enveloped single-stranded negative-sense RNA viruses such as measles, mumps, human parainfluenza, canine distemper, Hendra, and Nipah viruses, which cause a tremendous global health burden. The ability of paramyxoviral glycoproteins to merge viral and host membranes allows entry of the viral genome into host cells, as well as cell–cell fusion, an important contributor to disease progression. Recent molecular and structural advances in our understanding of the paramyxovirus membrane fusion machinery gave rise to various therapeutic approaches aiming at inhibiting viral infection, spread, and cytopathic effects. These therapeutic approaches include peptide mimics, antibodies, and small molecule inhibitors with various levels of success at inhibiting viral entry, increasing the potential of effective antiviral therapeutic development.

Introduction

The Paramyxoviridae family possesses some of the most infectious and lethal of the enveloped viruses, and although vaccines for a select few paramyxoviruses are available, outbreaks in unvaccinated and vaccinated individuals have increased. For example, measles (MeV) and mumps (MuV) viruses continue to pose a threat to public health, specifically in young children, despite the availability of a vaccine [1,2]. MeV in the genus Morbillivirus causes over 110 000 MeV cases per year [3,4]. The number of outbreaks due to MuV and MeV has also increased due to subpar vaccines, the waning of vaccine-induced immunity, and individuals choosing not to vaccinate themselves or their children [5,6]. Another Morbillivirus for which there is a vaccine and is of great veterinary concern is canine distemper virus (CDV), which causes a major health burden to a wide range of wildlife including dogs, lions, seals, and non-human primates [7-10]. Further, MuV in the genus Rubulavirus has had an increasing number of outbreaks since the early 2000s in both vaccinated and unvaccinated populations [11].

Additionally, there is the threat of emerging paramyxoviruses, for which there are no approved human treatments or vaccines, thus requiring urgent research and development. Such is the case of the deadly Henipaviruses (HNV), with many outbreaks in the past two decades. The initial Nipah virus (NiV) outbreak occurred in December 1998 in Malaysia, and had a mortality rate of ~40%, killing 105 of the 265 infected individuals [12,13]. Since then, there have been nearly annual outbreaks of NiV in Bangladesh [14-16]. Recently, NiV has been reported to have a mortality rate ranging from 40 to 100% depending on the outbreak [17,18,19^•]. For example, NiV claimed 21 of the 23 infected individuals during the 2018 outbreak in Kerala, India [20^•]. There are experimental treatments for NiV authorized for use during outbreaks. For example, ribavirin, a purine nucleoside analog capable of inhibiting viral RNA polymerases was used to treat NiV encephalitis during the 1998 and 2018 outbreaks. Ribavirin was used as a post-exposure prophylactic that reportedly successfully protected health care workers from contracting NiV [21,22,23^•]. The related HNVs, Hendra (HeV) and Mojiang virus, have also claimed several human lives [24-28]. Notably, the major host reservoir for the HNVs is the Pteropus genus of fruit bats which have a large geographic distribution worldwide from Australia, Southern Asia, and Africa [29,30]. This has likely led to increasing spillover events as human activity pushes into the fruit bat territory [31]. There are recent reports of seropositive bats in Brazil, underscoring the growing threat that HNVs pose to global public health [31,32].

MeV and MuV vaccines are archetypes for successful vaccine implementation and disease prevention. Nevertheless, many paramyxoviruses have no approved vaccines, leaving humans and livestock vulnerable [33]. Though there are several means of viral entry, regardless of the specific mechanims, a form of membrane-membrane fusion seems to be the common process by which most enveloped viruses enter naïve cells. Therefore, the development of antivirals targeting paramyxovirus entry is critical. Major hurdles in developing antivirals targeting paramyxoviral entry are the virus’ ability to mutate and the persistent emergence of novel strains. With improved vaccination strategies and effective antivirals, it may be possible that some paramyxoviruses will join smallpox and rinderpest (RPV) on the list of eradicated viruses [34,35].

Herein we review the major themes that have surfaced in the development of paramyxovirus fusion inhibitors. Including peptide, antibody, and small molecule approaches, which seek to interfere with crucial steps of the fusion process from initial receptor binding events to intermediate events in the membrane fusion and viral entry processes.

Paramyxovirus fusion

Paramyxoviruses require specific receptors on target cells for viral binding and entry to occur. Morbilliviruses such as measles, canine distemper, and peste des petits ruminants virus (PPRV) commonly use signaling lymphocytic activation molecule (SLAM) and nectin-4 receptors for entry into epithelial tissue [36-38]. The HNVs such as Ghana (GhV), NiV, HeV, and Cedar virus (CeV) can all utilize ephrinB2 and ephrinB3 as entry receptors, however, CeV can also use class A ephrins [39-45,46^••,47^•], and the receptor for Mojiang virus (MojV) is still unknown. Heparan sulfate has also been reported to function as an attachment factor that aids in the initial interactions of HNV glycoproteins to the target cell [48]. In the Avulavirus, Respirovirus, and Rubulavirus genera the HN glycoproteins require sialic acid-containing proteins as receptors for binding and entry, with sialic acid acting as the receptor [49-54].

Paramyxovirus fusion is coordinated by the action of two membrane glycoproteins, the receptor-binding protein (RBP) or attachment (G, H, HN) and fusion (F) glycoproteins. The RBPs are type II transmembrane (TM) proteins that form tetramers as dimers of dimers [55,56] and bind the host receptor initiating the fusion cascade by triggering the fusion glycoprotein (F) [57-61].

The trimeric paramyxovirus F glycoprotein is a type I TM protein and a class I fusion protein. The precursor F₀ protein must be proteolytically cleaved for F to be fusogenically active. HNV fusion proteins are not proteolytically cleaved as they are trafficked through the secretory pathway, and must be endocytosed to be fusogenically active [62]. HNV F₀ is endocytosed by Rab5-positive early endosomes to endosomal compartments, where F₀ is cleaved by cathepsin B/L or furin proteases and is then transported to the cell surface [63-67]. Fusogenically active F is composed of F1 and F² subunits linked by a disulfide bond [68]. The F of MeV, HPIV3, and Simian virus 5 (SV5) is proteolytically cleaved in the trans-Golgi network by furin [69,70^•,71]. The fusion proteins of Human metapneumovirus (HMPV), a former member of the Paramyxoviridae family now in the Pneumoviridae family, and Sendai virus (SeV) are matured into their fusogenically active form by extracellular proteases [72,73].

The structural core of F contains the heptad repeat (HR) regions HR1/A and HR2/B [74-76]. HR1/A is adjacent to the fusion peptide and HR2/B is located near the TM domain. The F2 subunit of F possesses the fusion peptide and the HR3 domain. HR1/A aids in the stabilization of the metastable prefusion state of F, the helical structure of the F-trimer [77,78]. HR1 transitions into an elongated helix, which develops into a trimeric coiled-coil stalk, causing F to be harpooned into the target membrane forming the pre-hairpin intermediate (PHI). Then membrane fusion occurs when HR1/A and HR2/B regions fold together to form a six-helix bundle (6HB). This action facilitates the merging of the viral and cellular membranes [79,80].

Paramyxoviral fusion begins when the RBP binds to its host receptor and undergoes conformational changes that trigger F to go from the pre-fusion state to the pre-hairpin intermediate (PHI) [81,82]. Such is in the case for Nipah fusion, where G interacts with F by both the stalk and globular head in a bidentate interaction [83]. Triggered F transitions from the pre-fusion to the PHI to the 6-helix bundle conformations, and the latter transition is thought to merge the cellular and virus membranes, allowing for the viral genome to enter the cell to begin the infection, or allowing cell-cell fusion during syncytia formation [84]. After F triggering, F and G are thought to dissociate, however, the mechanism(s) for this event is unknown. Additionally, it has been shown for HeV that F requires dissociation of the TM domain for fusion to proceed [85^••]. In summary, paramyxoviral membrane fusion occurs primarily at the plasma membrane and is a pH-independent process requiring the presence of F and the RBP on the effector viral or cellular membrane, and a receptor on the target membrane to initiate the fusion cascade [86,87]. Cell-cell fusion (syncytium formation) occurs when F and RBP are expressed on the surface of an infected cell and interact with the host receptor on an adjacent cell. The formation of giant multi-nucleated cells is a hallmark of paramyxoviral fusion. This is thought by many as a form of viral genome transmission free of viral particles [42,88-90].

Currently, there are five models of F triggering ranging from G promptly dissociating from F after receptor binding, to high-order oligomerization of F with G after receptor binding [80]. Furthermore, the minimal and maximal ratios of F to G necessary to execute paramyxoviral membrane fusion are unknown. However, a recent study by Xu et al. suggests that a greater proportion of NiV F hexamers may be optimal for NiV membrane fusion [91].

Paramyxovirus pathology

Paramyxovirus infections can elicit a wide range of symptoms. Newcastle disease virus (NDV), also known as avian paramyxovirus 1, in the genus Avulavirus, presents clinical signs in birds including diarrhea, lesions in the brain, kidney, and pancreas, head tremors, and ataxia [92]. NDV has been responsible for significant economic loss in the poultry industry [93].

HNV infections severely affect the central nervous system. Humans infected with NiV developed endothelial syncytium formation, vasculitis, and thrombosis [94]. Pigs are susceptible to NiV infection and display respiratory distress that is accompanied by a ‘barking-cough’ [95]. However, pigs are not as severely infected with NiV as humans, though pigs can spread the virus easily amongst themselves. During the 1998 outbreak, over a million pigs were culled to stop the spread of NiV [18,96]. HeV, closely related to NiV, is deadly to both horses and humans. The initial HeV outbreak occurred in 1994 in Queensland, Australia and resulted in the death of two humans and 14 horses. The first human fatality suffered from respiratory hemorrhaging and edema associated with syncytia formation in the lungs, in addition to headaches, sore throat, a stiff neck, vomiting, and seizures, before succumbing to the infection. The second fatality from HeV was an individual that had contracted HeV a year earlier. Giant-multinucleated cells were found in the viscera and brain of this individual at autopsy [97,98].

HPIV infections are usually self-limited and cause bronchiolitis, croup, and pneumonia in children. HPIV 1–3 are responsible for a high mortality rate in immunocompromised adults and young children [99-101]. In the U.S., HPIV related pneumonia was responsible for an annual 10 100 hospitalizations in children younger than five years old [102]. MuV causes painful swelling of the parotid gland, scrotum in males, and breasts in females. Individ-uals infected with MuV also experience nausea, vomiting, headaches, and fever [103]. MuV can also cause a reduction in testes size and infertility in men and premature menopause and infertility in women. However, cases of infertility due to MuV are rare [104,105^••,106].

In the Morbillivirus genus, MeV and its persistent cell–cell only variant, Subacute sclerosing panencephalitis (SSPE), can cause life-threatening infections in young children. Following the initial MeV infection, after six years on average, when SSPE develops it causes demyelination within the central nervous system, leading to increasingly severe neurological degeneration culminating in death within four years [107,108]. Additionally, MeV elicits acute immune suppression by infecting memory B, T, and plasma cells, virtually eliminating the antibody repertoire against other pathogens [109^•]. Clinical signs of CDV infections are diarrhea, vomiting, respiratory distress, and neurological symptoms, including ataxia and convulsions [110]. CDV infection in the CNS can cause demyelination in the white matter and syncytium formation, leading to neuropathology [111-113]. Feline morbillivirus is a newly emerging virus thought to cause chronic kidney disease and tubular intestinal nephritis in cats [114,115^•]. Further, atlantic salmon can be infected by Atlantic salmon paramyxovirus from the Aquaparamyxovirus genus. Infection in the gills induces respiratory distress and a 40% mortality rate in salmon farms [116]. Peste des petits ruminants virus (PPRV), a highly contagious Morbillivirus that commonly infects goats and sheep, causing necrotic lesions in epithelial cells and lymphoid tissue [117]. PPRV also causes fever, ocular, and nasal discharge, and has a mortality rate of 70–80% in infected ruminants [118].

Fusion protein peptide antivirals

Paramyxoviral fusion inhibitor peptides target-specific regions of F to arrest membrane fusion. Most of the inhibitory peptides are sequences derived from the HR regions, however, the TM domain and the leucine zipper regions have also been targeted. Most peptide-based inhibitors are mimics of the heptad repeat regions of the fusion protein, which upon binding prevent the formation of the six-helix bundle (6HB) and arrest fusion at an intermediate step. This is the mechanism of action for the class I fusion peptide enfuvirtide which is a 36 amino acid (aa) peptide of the gp41 HR2 domain that inhibits HIV entry. However, native sequence-based peptides have not proven sufficient to successfully inhibit in vivo, prompting further enhancement of inhibitory activity by altering the peptide sequence. These peptide mimics have been remarkably effective as inhibitors of viral entry in vitro and in vivo [119].

The efficiency and bioavailability of these peptide mimics can be improved via various modifications including conjugation to hydrophobic molecules such as cholesterol, which greatly increases membrane intercalation and significantly reduce the amount of self-aggregation [120,121]. This same approach has also been successful in inhibiting in vitro and in vivo infections of HMPV, NiV, and MeV [122-125]. Notably, several peptides have shown the ability to impede fusion across different viral species, families, and orders. The efficacy of peptide-induced inhibition can be enhanced by altering peptide length and/or sequence, and inclusion of hydrophobic moieties to enhance binding to the fusion protein. For example, shorter peptides of the HR2/B region of HPIV2 and HPIV3 showed only reduced levels of infectivity but longer peptides completely inhibited viral entry and spread [126].

Interestingly, an HPIV-HRC peptide yielded a greater reduction in fusion toward live NiV and HeV than that of the HRC peptides of NiV and HeV peptides, respectively. This was thought to be due to the inverse correlation between F-triggering kinetics and peptide efficacy since NiV and HeV have longer fusion protein activation time than HPIV3 [127-130]. To probe the heterotypic anti-HeV efficacy of HPIV3 peptides, HeV-HPIV3 homotypic and chimeric peptides revealed that the HPIV3-HRC peptide interacts with the coiled-coil trimer of HeV-F, thereby blocking the formation of the 6HB [128]. Furthermore, the inhibitory effect of the HPIV3-HRC peptide toward HPIV3 entry was improved with the addition of a cholesterol moiety to the C-terminus or N-terminus of the peptide. However, only the HPIV3-HRC peptides that were cholesterol tagged on the C-terminus had improved inhibition toward HeV or NiV pseudotyped virus and live virus entry [131].

Sequence optimization of HPIV3-HRC yielded a peptide, ‘VIKI’, with substitutions (E459V, A463I, Q479K, K480I) that improved inter-helical binding. The VIKI peptide with a cholesterol moiety and a PEG spacer at the C-terminus effectively inhibited infection of HPIV3 and NiV at low nanomolar concentrations in vitro. Prophylactic administration of VIKI-PEG₄-chol two days before infection completely protected golden hamsters for up to 21 days after challenge with 100 × LD₅₀ of NiV. Simultaneous infection of NiV in golden hamsters with VIKI-PEG₄-chol treatment garnered a survival rate of 80%, while treatment beginning 2 days after infection yielded a survival rate of 40% [130]. Further, a cholesterol-tagged dimer of HPIV3-HRC with a PEGylated linker effectively inhibited viral entry of both HPIV3 and NiV and improved the survival of NiV infected hamsters when intraperitoneally administered [132]. The HPIV3-HRC peptide was further improved with the conjugation of a PEG₂₄-cholesterol linker at the C-terminus, which interestingly greatly improved inhibition of NiV, while the N-terminally conjugated peptide preferentially inhibited HPIV3 [133]. Moreover, HPIV3-HRC with ‘VIKI’ substitutions conjugated to dPEG₄ and tocopherol (toco) at the C-terminus produced a lipopeptide with improved protease resistance. The VIKI-dPEG4-toco lipopeptide yielded a survival rate of 50% in Syrian golden hamsters with only 3 administrations after a 100 × LD₅₀ inoculation of NiV. The VIKI-dPEG4-toco peptide was also tested in African green monkeys and resulted in a survival rate of 33% compared to the untreated monkeys which all succumbed to NiV infection [134]. Another HPIV3 36-mer peptide, ‘VIQKI’ (E459V, A463I, D466Q, Q479K, K480I), with substitutions to enhance interactions with the fusion protein’s HRN domain to the PHI, inhibited fusion of HPIV3 and respiratory syncytial virus (RSV), a former member of the Paramyxoviridae family now placed in the Pneumoviridae family [135]. VIQKI peptides with either an I454F or I456F substitution or both had improved antiviral efficacy toward HPIV3 and RSV due to improved binding of the trimeric core [136]. Several versions of HPIV3-HRC peptides inhibit MeV-induced cell–cell fusion at early time points, but could not reduce MeV titers in vivo [137]. Finally, single residue substitutions of α-to-β amino acids within the N-terminal region of VIQKI yielded peptides with comparable inhibitory activity compared to wt VIQKI but induced differing interactions with HPIV3-HRN [138^••].

Inhibitory peptides can also block Rubulavirus entry. Peptides derived from the HRA (N-1) and HRB (C-1) regions of SV5 both effectively inhibited syncytium formation during incubation in cells expressing F and HN. However, the N-1 peptide inhibited fusion only at an early fusion intermediate stage; after receptor binding but before F activation. The C-1 peptide inhibited fusion by binding to the PHI before 6HB formation occurred [139].

A dimerized sequence of MeV-HRC conjugated to cholesterol (MeV-HRC4) inhibited MeV entry and spread in Vero cells. Mice prophylactically treated with MeV-HRC4 one day before infection and daily treatments for 1 week after infection protected mice from MeV-induced encephalitis. The brains of infected, non-treated mice revealed the presence of MeV nucleoprotein while mice treated MeV-HRC4 had no detectable viral antigen [140]. A MeV-HR2 peptide effectively inhibited infection of MeV into Vero cells, and interestingly an HPIV3-HRC peptide inhibited MeV entry into SLAM bearing cells [122,137]. Moreover, intranasal delivery of MeV-HRC4 protected mice better than subcutaneous injection by 80% compared to 60%, respectively [137]. Dimerized MeV-HRC peptides conjugated to toco or cholesterol and significantly reduced infection in cotton rats. MeV-HRC conjugated to toco remained in the lungs to a greater extent than the cholesterol conjugated peptide which localized more in the brain. In general, anti-MeV peptides inhibited fusion better in cells expressing receptor CD150 than nectin-4 [137,141].

Furthermore, a MeV-HRC peptide conjugated to 25-hydroxycholesterol demonstrated significant inhibitory activity in vitro [142]. SSPE, a persistent cell-cell only variant of the measles virus, [108] can also be inhibited with peptides generated from the MeV Edmonton strain HR2 domain, which significantly inhibited MeV plaque formation and improved the survival of SSPE infected mice to 67% [143]. A peptide of the α-helical region of MeV F (455–490) effectively inhibited syncytium formation in MeV and CDV infections and significantly reduced MeV entry into Vero cells [119].

Interestingly, peptides from Marek’s disease virus (MDV), from the family Herpesviridae, had inhibitory activity against NDV. To enhance cross-reactivity toward NDV and MDV, these peptides sequences from gH or gB proteins were linked to the HR region of NDV. Interestingly, the tandem peptide constructs had superior inhibitory activity than that of their monomeric peptides [144].

The F TM domain has also been a target for inhibitory fusion peptides. HeV-F TM domain peptides altered F stability, the ability of F to form homo-trimers, expression of F, and significantly inhibited syncytium formation but did not alter F’s ability to be cleaved into its active fusogenic form. This approach was also successful in reducing PIV5 infection with a PIV5 TM domain peptide [145^•]. A peptide derived from the Respirovirus, SeV F protein sequence (473–495) inhibited hemolysis of erythrocytes at a low micromolar concentration after SeV had attached to the cell [146].

Anti-paramyxovirus antibodies

Another successful approach to paramyxovirus entry inhibition is antibody therapy. Similar to the approach used for early HIV antivirals, neutralizing antibodies effectively bind regions of the viral glycoproteins necessary for fusion [147].

Rabbit monoclonal antibody (mAb) mAb66 binds to NiV-F and inhibits viral entry. Interestingly, antibody neutralization by mAb66 is unhindered by the presence of F glycosylation. However, it was observed that mAb66 has a better binding affinity toward NiV-F than HeV-F [61,148,149•]. The antigen-binding fragment (Fab) of mAb66 was revealed to bind within domain III (DIII) of the globular head of NiV-F. Likewise, mAb36 had a similar binding site to mAb66 but bound only HeV-F. Interestingly, 2 aa substitutions, Q70K + S74T, to NiV-F conferred neutralizing capabilities to mAb36 equivalent to HeV-F. Additionally, another polyclonal antibody, pAb835, also targets the F globular head in neutralizing NiV and HeV, supporting the hypothesis that the immune system commonly targets the apex of the HNV-F globular head for neutralization [149^•]. Murine antibodies 1B2, 7G9, and 3D7 inhibited pseudotyped NiV entry by binding to NiV-G. However, these mAbs did not possess cross-reactivity to the closely related HeV [150^•].

Human monoclonal antibodies m102 and m102.4 are reactive towards the ephrinB2/ephrinB3 binding sites on NiV-G and HeV-G [151]. Treatment with m102.4 10 hours post-challenge protected ferrets infected with NiV [152]. African green monkeys were completely protected from illness after HeV challenge when m102.4 was administered 3 days post-challenge. Furthermore, m102.4 can neutralize HeV and the Malaysian (M) and Bangladesh (B) strains of NiV and completely protected African green monkeys infected with NiV (M) 5 days and NiV (B) 3 dpi [153-156].

Murine 5B3 and humanized h5B3.1 antibodies inhibited HNV fusion binding to DIII of the globular head thereby locking F in the prefusion state. However, an escapemutant was generated after only three passages of live NiV with 5B3. The NiV-F escapemutant possessed a K55E mutation and was completely uninhibited by antibody 5B3, resulting in wild-type levels of fusion [157^•].

The human monoclonal antibodies, 54G10, 17E10 binding to antigenic site IV on HMPV-F and RSV-F, while MPV364 bound to antigen site III of only HMPV-F [158,159^••]. Mice prophylactically treated with 54G10 or MPV364 had reduced viral titers in the lungs compared to control mice infected with HMPV [159^••,160]. Anti-HMPV-F Fab DS7 can neutralize genotypes A1, A2, and B1. Electron microscopy revealed that DS7 binds the DI and DII domains of the HMPV-F head. DS7 reduced HPMV plaque formation by 60% and significantly reduced viral titers in the lungs of cotton rats when administered 3 days after infection [161,162].

A murine mAb m77.4 that binds to the prefusion state of MeV F was constructed into a single-chained variable fragment (scFV) and inhibited viral fusion at low nanomolar concentrations. Prophylactic administration of the scFV into cotton rats induced significantly lower viral titers in the lungs after infecting with MeV [163].

An anti-HN mAb M1-1A inhibited HPIV2 induced cell-cell fusion by binding to the HN stalk at residues N83 and K91 without altering HA or NA activity. Inhibition of virus-cell fusion by M1-1A occurred after adsorption but before membrane merging [164-167]. Synthetic antibodies (sAb) have also been effective at neutralizing PIV5 entry. SAb 5D inhibited PIV5 fusion by binding to the N-terminal domain thereby stabilizing the prefusion state of F or by blocking F-HN interactions. Anti-PIV5 HN sAbs 1H and 4E bind to the HN stalk region (1–117) while sAb F4 binds to the HN NA domain and each is able to inhibit virus-cell fusion and syncytia formation [168,169].

Though anti-paramyxovirus antibodies are effective at inhibiting viral entry there still exists the possibility of resistance and escape-mutants being generated during their use [60,157^•,170]. In addition, the cost and levels of antibodies needed to prevent or cure paramyxoviral infections may prohibit their extensive use in the case of a global pandemic.

Small-molecule antivirals

Several small-molecule antivirals have been discovered that target domains conformationally important in paramyxoviral fusion. For example, small non-peptide mimics, CSC7 and CSC11, induce premature activation of HPIV3 F. Treatment of CSC7 and CSC11 at 3 mM in HPIV3 infected HAE cells reduced viral titers by ~75% and ~100%, respectively. These compounds did not inhibit receptor binding but were found to prematurely activate the HN/F fusion mechanism before receptor binding [171]. Using conformational antibodies against HPIV3 to measure the activity of CSC11 allowed for the development of improved CSC11 derivatives. One such compound, CM9, had significantly improved inhibitory activity (IC50 = 920 μM) compared to CSC11 (IC50 = 2000 μM). A 3 mM treatment of CM9 completely inhibited HPIV3 infection of both laboratory and clinical strains. Structural antibody analysis revealed that CM9 induced F to transition into the 6HB state [172^••]. Similar to CM9, compound PAC-0366 was able to trigger HPIV3 F in the absence of receptor binding with an improved IC50 by 100-fold compared to CM9 [173].

There have been reports of antiviral compounds that are effective against the influenza virus being excellent inhibitors of paramyxoviral entry. One such instance is Zanamivir, a neuraminidase inhibitor against HPIV3 infections. The pre-incubation of HPIV3 with Zanamivir during adsorption resulted in a decrease in plaque numbers, but not the size of the plaques. Zanamivir has an affinity toward the NA active site and sites where HN interacts to sialic acid receptor, which inhibts HPIV3 infectivity [174]. However, Zanamivir enhanced the HN receptor avidity in NDV entry, due to the occupation of the binding site I by Zanamivir, which activated the HN binding site II inducing greater receptor avidity [175]. Nevertheless, Zanamivir is an effective inhibitor of HPIV and NDV NA activity, but NDV is resistant to Zanamivir’s inhibitory effects on receptor binding and viral entry [176].

Additionally, glycan-mediated inhibition of fusion with griffithsin (GRFT), a high-mannose oligosaccharide binding lectin, was able to block viral entry and spread of NiV. GRT inhibits NiV-induced membrane fusion by interacting with NiV-G, thereby impeding virus-cell interactions. Trimeric GRFT (3 mG) had superior antiviral potency than monomeric GRFT. Prophylactic administration of 3 mG and an oxidation-resistant GRFT (Q-GRFT) protected hamsters from lethal intranasal challenge with NiV. However, Q-GRFT induced a greater survival rate (35%) than 3 mG (15%) [177].

A novel antiviral method aimed at reducing viral binding used heparan sulfate (HS) to block receptor binding of HMPV. Infection was reduced in BEAS-2B cells when pretreated with sulfated iota-carrageenan before infection with HMPV. This reduction in infectivity was due to iota-carrageenan competing with HS binding sites on HMPV-F. Highly sulfated K5 polysaccharide derivatives that possess a heparan-like structure could also inhibit the ability for HMPV to bind HS [178]. Additionally, the process of NiV particles binding to cells could be inhibited or displaced when cells are pretreated with heparin [48]. Furthermore, an HRA2 analog peptide expressed in Nicotiana tabacum plants was an effective inhibitor of HMPV binding to HS [179].

Recently, Presatovir underwent phase II clinical trials as a fusion inhibitor for RSV [180]. In healthy adults Presatovir successfully reduced viral load and symptom score compared to patients receiving a placebo [181]. Presatovir inhibits RSV induced membrane fusion by binding to the prefusion state of RSV-F thereby preventing conformational changes necessary for membrane fusion [182].

Additionally, small compounds that inhibit MeV infections have been explored. For example, anti-MeV compounds oxazole 1 (OX-1) and its improved variant AM-4 were originally designed to bind MeV-F near the FP. However, despite AM-4’s high inhibitory activity it had a short half-life [183,184]. Nevertheless, structural optimization of AM-4 led to the development of a more stable compound, AS-48, effectively inhibited several MeV genotypes as well as CDV and RPV. Both AS-48 and FIP, a fusion inhibitor peptide, bind to MeV-F in the N-terminal portion of HRB where the head and stalk connect. This protein-inhibitor interaction locks MeV-F in the prefusion state by linking individual F protomers at the hydrophobic pocket, which impedes conformational changes [185]. Interestingly, mutations to the MeV-F HRB region granted MeV resistance to AS-48 and FIP without negatively affecting membrane fusion. However, after several passages of the MeV without the selective pressure of FIP, MeV-F reverted to the wild-type sequence. This suggests that the beneficial MeV-F mutations against FIP came as a detriment to viral fitness [186-188].

Some of the most effective broad-spectrum antivirals are the LJ and JL series. LJ001 is a lipophilic thiazolidine derivative requiring light for activation and, in the presence of oxygen, produces singlet oxygen (¹O₂) free radicals. ¹O₂-induced allylic hydroxylation in unsaturated fatty acid chains stimulates trans-isomerization and the addition of a hydroxyl group in the lipid bilayer. LJ001 inserts into the viral membrane and by singlet oxygen (¹O₂) mechanism causes an increase in lipid packing density, thereby reduce membrane fluidity, which ultimately reduces virus-cell fusion. LJ001 has shown to be effective at reducing the viral spread between fish in an aquaculture environment. LJ001 is effective at inhibiting a variety of enveloped viruses: arenaviruses, bunyaviruses, filoviruses, flaviviruses, HIV-1, Influenza A, paramyxoviruses, and poxviruses. Treatment with LJ001 leaves the virus intact but inhibits the viral particle-cell fusion [189,190].

The development of small compound broad-spectrum antivirals such as LJ001 has become an objective of much research amidst the COVID-19 pandemic. Recent studies have aimed to understand the mechanism by which the SARS-CoV-2 virus becomes fusogenically active, and it has been the elucidation of this mechanism that has provided much insight into potential targets for broad-spectrum inhibition of enveloped viruses, notably paramyxoviruses and coronaviruses. For SARS-CoV-2 to invade host cells, the activation of its Spike protein (S protein) is a requirement that is ultimately achieved through the proteolytic cleavage of S by transmembrane protease serine 2 (TMPRSS2), a type II transmembrane serine protease (TTSP), into an S1 and S2 subunit. The S1 subunit binds to the angiotensin converting enzyme 2 (ACE2) receptor of the cell membrane, which facilitates the attachment of SARS-CoV-2 to the surface of a host cell. The S2 subunit drives the fusion of viral and cellular membranes that results in viral infection [191,192]. Considering its role in priming the S protein for viral entry, the inhibition of TMPRSS2 is a prime target for the development of antivirals for SARS-CoV-2. In vitro experimentation with camostat mesylate, a clinically proven serine protease inhibitor that is active against TMPRSS2, demonstrated reduced SARS-CoV-2 entry and infection in a Caco-2 cell line [193^••]. In vivo experiments with SARS-CoV-2 are still ongoing, in which the efficacy of camostat mesylate, both alone and in combination with other proposed antivirals, will be assessed [194]. These results offer a potential broad-spectrum application when considering the similar role transmembrane serine proteases have been proven to fulfill in the cleavage of hemagglutinin (HA) protein in influenza A virus (IAV) and the fusion protein (F protein) in HPIVs and other paramyxoviruses. Protease inhibitors such as the new broad-spectrum N-0385 may prove useful across viral families [195^••]. Camostat mesylate demonstrated a reduction in replication of IAV in human tracheal epithelial cells as a result of a reduction in proteolytic cleavage of HA by TMPRSS2 [196]. To date, there have not been any conclusive studies that have proven the effect of serine protease inhibitors on paramyxovirus entry. This is believed to be a result of a lack of understanding behind the main proteases responsible for pathogenesis. Nonetheless, the evidence provided by present studies indicates that serine protease inhibitors have an immense potential for use as broad-spectrum antivirals. Further elucidation of viral entry mechanisms and the roles of TTSPs will be crucial to assessing this possibility in the future.

Conclusions

Great strides have been made in the development of potential antivirals towards paramyxovirus fusion as shown in Figure 1. However, given the limited options of vaccines and antivirals, more prophylactic options are needed. The U.S. has had great success in reducing the outbreaks of MeV with an effective vaccine and schools mandated vaccine requirements [197^•]. The same cannot be said globally where suboptimal vaccines have left a large population vulnerable to MeV infections. Furthermore, recent serological screenings worldwide of bats and rodents have demonstrated that there is a continuously growing number of viruses that have the potential to causes serious health and economic burden. This underscores the urgency for the development of new antivirals. Since the advent of the first successful peptidomimetic HIV-1 antiviral T-20 (enfuvirtide), there have been a plethora of homologous paramyxovirus peptides capable of inhibiting viral entry. Fortunately, if viruses share enough fusion protein structural homology, a peptide or small molecule for one virus may be effective for another. There have been a few successful attempts at producing broad-spectrum antivirals for enveloped viruses [198]. However, although these antiviral approaches have been successful, they all face the hurdle of the appearance of resistance mutations. To combat the constant emergence of new paramyxoviruses, the future of paramyxovirus and all enveloped viruses’ fusion inhibitors may include having a common target unable to mutate, such as the viral membrane, such as the LJ and JL class of antivirals that protected fish from the horizontal transmission [199]. However, a major downside with these antivirals is the need for light for activation. Nevertheless, they serve as an archetypal antiviral method that should be further explored. An ideal antiviral would be able to target the viral membrane and completely halt the ability of the viral membrane to fuse with the cell membrane. Furthermore, such antivirals should leave the viral proteins and genetic material unaltered. This could potentially allow for large scale inactivation of viruses that may aid inactivated-virus vaccine development. With the viral glycoproteins unaffected, the antigenic sites would remain relatively more intact, allowing for greater immune responses post-challenge.

Figure 1.

Summary of Paramyxovirus antiviral target locations within the membrane fusion cascade model. Green bars indicate anti-paramyxovirus antibody epitopes, Red bars show peptide targets, and blue bars point to small-molecule antiviral interaction sites. PHI = Pre-hairpin Intermediate, 6HB = 6 Helix Bundle.