The basis of the seemingly successful repurposing of IVM is rooted in the identification of important components encoded by the viral genome. The SARS-CoV-2 viral genome encodes non-structural, structural, and accessory proteins. Its positive mRNA strand is translated within the host cell in order to, first, produce its own replication machinery, and second, to produce the structural components required to house viral progeny (10). Two-thirds of the genome code for two large polyproteins, pp1a and pp1ab. Once formed, the polyproteins are subsequently cleaved into 16 individual non-structural proteins (nsps), which primarily provide enzymatic activity (11). Three nsps (1–3) are cleaved by papain-like proteases (PLpro), which itself is localized within nsp3, and the rest are cleaved by the main protease (3C-like protease, 3CLpro) on nsp5 (1). As such, translation of the viral PLpro and 3CLpro are essential for efficient reproduction of the virus. Once the nsps are available, they cooperatively form the replicase-transcriptase complexes (RTCs), which are required for the production of new virions (12). Some nsps (3,4 and 6) induce the development of double membranes from the endoplasmic reticulum (E.R.), Golgi apparatus (G.A.) or the ER-Golgi intermediate compartment (ERGIC), which serve as foci for viral genesis (12). Collectively, the rest of the nsps in the RTC include RNA polymerase, helicase, exoribonuclease, and methyltransferase, among many others. The exact mechanism of replicating its own genome is still under investigation. However, it is understood that negative-sense intermediates are initially formed and then serve as templates for reproducing both genomic and sub-genomic positive-sense RNAs (13). A potential model for the RNA replication in SARS-CoV-2 has been postulated and it is based on homologous proteins in SARS-CoV-1 (10).

Structural proteins are highly conserved among the various genera of coronaviruses. They include the spike protein (S), the envelope protein (E), the nucleocapsid protein (N) and the membrane protein (M). Once the structural proteins are synthesized, and the viral RNA is reproduced, the S, M and E become embedded within the previously formed double membranes from the host E.R. and eventually reach ERGIC. Meanwhile, the N protein which is tethered to the newly formed genome ‘delivers’ this RNA into S-M-E-embedded ERGIC membrane, leading to the formation of ‘pockets’ which eventually seal off into new virions (1). The interaction of N with the 3’-end of the viral genome is mediated via nsp3 (14), the largest subunit of the RTC. The nsp3 acidic ubiquitin-like N terminal domain (UbI1) binds to a serine- and arginine-rich domain in the N protein, thereby anchoring the viral genome to the RTC in order to facilitate RNA replication and, importantly, to eventually ensure the localization of the newly synthesized genome within the viral envelope (Hurst, Koetzner, & Masters, 2013). Ultimately, the N protein is incorporated in the RNA helical structure, which underlies the envelope (15). Overall, the N protein enhances coronavirus transcription, interacts with the viral genome and with M in the viral envelope. Notably, inhibition of N was shown to greatly suppress viral replication, suggesting it is an essential factor in efficient virion production (14, 15). Interestingly, N is the highest expressed protein in infected cells, further corroborating its importance in the viral life cycle (15).

Although RNA replication and translation occur in the cytosol, nuclear access is a key event in the infectious cycle of several viruses, including coronaviruses (1, 8). However, the entry of proteins into the nucleus is a tightly regulated process. To evade this limiting barrier, some viral proteins exploit the importin (IMP) superfamily of nuclear transporters to gain nuclear access (16). Nucleocytoplasmic trafficking is mediated via transmembrane nuclear pore complexes (NPCs) in the nucleus, composed of nucleoporin (NPR) subunits. A major class of NPRs known as FG-NPRs are distributed throughout the NPCs and enable nucleocytoplasmic transport due to their interaction with IMP transporters (17). The major IMP classes include IMPα and IMPβ. Nuclear import is mainly mediated either by IMPβs or by heterodimers of IMPα/IMPβ1 (17–19). For cytosolic protein cargo destined for nuclear import, IMPs, particularly IMPα proteins, recognize nuclear localization signals (NLS) on target cargo proteins, whereas IMPβ facilitates the actual transport via the NPCs (18). Efficient target binding to the IMPs/EXPs is further supported by Ran, the small monomeric GTPase (20). Active Ran causes dissociation of IMPβ from the importin/NLS-protein complex, releasing its tethered cargo into the nucleus (21). Thus, for a potential SARS-CoV-2 protein to reach the nucleus, it must contain an NLS, properly interact with IMP proteins and Ran must be activated.

As it happens, the SARS-CoV-2 N contains NLS motifs. Of great significance is the finding that NLS regions in the N gene of SARS-CoV viruses are highly variable compared to the NLS of other coronavirus clades (22). Importantly, these changes occurred during the recent evolution of the highly pathogenic coronavirus clades- including SARS-CoV-2 (22). Incidentally, the numerous nucleotide insertions and deletions within the NLS are associated with enhanced nuclear translocation. Three NLS motifs have been detected on the N of SARS-CoV-2, SARS-CoV, MERS-CoV and seasonal coronaviruses. Uniquely, as a result of the nucleotide variations found in SARS-CoV-2 and SARS-CoV-1, all three NLS motifs contain a distinctly higher overall positive charge among the peptides compared to the less virulent coronaviruses. The higher positive charge of NLS renders the entire N protein also more positively charged and subsequently enhances its efficacy (23). It has been previously corroborated in animal studies that the enhanced translocation of viral Ns to the nucleus results in more severe pathogenicity (24). Therefore, it is possible that these more positively charged Ns, which are characteristic of SARS-CoV-2, may be partially responsible for the associated detrimental effects.

It was previously shown that viral proteins that enter the nucleus might suppress host genes related to the anti-viral response, leading ultimately to increased pathogenicity (25). This may also be the case with SARS-CoV-2, as in vitro studies indicated that the SARS-CoV-2 NP could interact with dsDNA, possibly due to its high positive charge and the negative charge of DNA (26). Although the exact activity of the SARS-CoV-2 N within the nucleus has not been fully characterized, previous examination of several coronavirus Ns can offer insight (24).

The N of the coronavirus infectious bronchitis virus (IBV) was detected not only in the cytoplasm but also within the nucleolus. Nucleolus targeting was also shown with the SARS-CoV-1 N (27). It is important to note that the presence of N in the nucleus was indispensable for the replication of IBV, highlighting that cytosolic activity was not sufficient. In another related coronavirus, mouse hepatitis virus (MHV), nuclear proteins were also implicated in its replication. MHV N was specifically detected in the nucleolus, which itself is formed during interphase of the cell cycle and allows formation of ribosomal RNA (rRNA) and ribosomal subunits. The reason for N targeting of the nucleolus is not entirely understood. However, it is possible that N associates with rRNAs, in order to ‘reserve’ their use for translation of sub-genomic RNA. It was also shown in vitro that N transfection into cells resulted in multi-nucleate cells, indicating the delay of cytokinesis (24). This would provide favorable and prolonged conditions for the virus intracellularly to continue to synthesize its genome and sub-genome, translate its proteins and enable sufficient virion packaging. Moreover, N is proposed to dampen the host cell’s antiviral transcriptional response within the nucleus (8). Nevertheless, confirming the presence SARS-CoV-2 N in the nucleolus and understanding its role would elucidate the pathogenicity of this virus.

N is an essential component of newly formed virions as it ensures a proper ‘delivery’ of the replicated viral RNA genome within the developing envelope (28, 29). Moreover, it is essential for proper viral RNA dependent RNA polymerase activity, as demonstrated in Influenza A (29). As such, targeting the activity of N would offer a potent antiviral activity against SARS-CoV-2. In fact, N was shown to be an effective anti-viral target against Influenza A. One of the useful properties of N is its numerous binding sites, which have been shown to accommodate various drugs (29, 30). For example, compounds which can target the tail-loop binding pocket abrogate N oligomerization, while the compound F66 binds to the RNA-binding groove of the protein and is associated with improved survival in animal models infected with Influenza A (29). Figure 1 illustrates how the N of SARS-CoV-2 facilitates virus replication and mitigates the host cell response, thus further strengthening its position as a promising target of anti-viral drugs.

IVM was originally discovered from organisms that were isolated from soil samples collected from the woods nearby to Kitasato Institute in Kawana, Japan. Fermentation products released by a bacterium from the soil, which was later classified as Streptomyces acermitilis, appeared to exhibit antiparasitic activity (specifically against Nematospiroides dubius). Purification and isolation of the bioactive compounds showed naturally occurring macrocyclic lactones, and these were subsequently named avermectins. Avermectins are made up of four compounds, which exist as two variants: A₁, A₂, B₁, and B₂. Variants ‘A’ and ‘B’ indicate the presence of methoxy or hydroxyl groups, respectively, at the C5 position. Number ‘1’ describes the double bond between C₂₂ and C₂₃. On the other hand, number ‘2’ indicates the presence of hydrogen at C₂₂ and a hydroxyl group at C₂₃. B₁ avermectins were proven to be most active on oral administration, and on this basis, IVM was chemically derived. IVM contains an 80:20 combination of 22,23-dihydro-acvermectin B_1a and 22,23-dihydro-avermectin B_1b. Its antiparasitic effects are primarily caused by high-affinity irreversible binding to glutamate-gated chloride (Cl^-) channels located on nerve and muscle cells of nematode, which leads to hyperpolarization (9, 31). Ultimately, the increased permeability to Cl^- results in paralysis and death of the nematode (31).

As of yet, IVM has treated hundreds of millions of people with onchocerciasis, most commonly given at 150-200 μg/kg of body weight for one dose initially, and repeated at 6-12 monthly intervals as appropriate (32). Its use extends to a broad spectrum of parasitic nematodes on both oral and parenteral administration, and is also effective against arthropods, including lice (33).

Importantly, IVM was approved by the FDA for human use in 1987 (34). Its low toxicity and safety are attributed to the fact that its human target receptors are ‘secluded’ in the CNS, and IVM has not been shown to cross the blood-brain barrier. In addition, IVM displays a 100-fold greater affinity for parasitic Cl^- channels compared to the human homologs (35). Moreover, severe detrimental effects in humans were shown only in those who over-dosed using approximately 15.4 mg/kg body weight IVM, which is 77 times above the prescribed dose. This corroborates the advantage of repurposing drugs, as these medications have already been tested arduously and extensively to confirm their efficacy and safety, thereby decreasing the transit time from shelf to intake.

The potential of IVM as an inhibitor of nuclear transport of viral proteins was initially suggested in 2011. Initially, Wagstaff and colleagues screened for nuclear import inhibitors, which block the interaction between IMPs and potential target cellular proteins (21). They randomly selected 480 compounds from LOPAC¹²⁸⁰ (Library of Pharmacologically Active Compounds; Sigma, St. Louis, MO). IVM surfaced as a drug that generally inhibits IMP activity (21). A year later, they confirmed that this apparent activity of IVM also inhibits nuclear transport of viral proteins HIV and Dengue virus in HeLa cells (36). Specifically, it was shown that GFP-tagged IMP was significantly reduced in the nucleus of HeLa cells after 3 hrs. of co-incubation with IVM (36). Moreover, the effect was unique to IMPα/β interactions and did not affect proteins bound only to IMPβ1. The importance of blocking the nuclear import of viral proteins emerged when it was later shown that IVM also prevented replication of HIV (36). As such, it surfaced as a possible repurposed drug, capable of preventing viral cargo from interacting with IMPα/β for nuclear import, with the potential to result in viral ‘death’ (21, 36).

Soon after, the effect of IVM against the nuclear import of viral proteins was further validated. For example, IVM prevented nuclear translocation of nsp5 in Dengue virus, West Nile virus, and influenza and inhibited transport of large tumor antigen (T-ag) in simian virus (25, 37, 38). The Wagstaff et al., 2011 and 2012 studies were pivotal in providing much of the initial rationale for the recent consideration of IVM as a SARS-CoV-2 antiviral agent. In fact, it was the same researchers who nine years later, in 2020, demonstrated that IVM inhibits SARS-CoV-2 in vitro replication (8).

Given its efficacy in inhibiting nuclear import of other viral proteins, the anti-viral effect of IVM against SARS-CoV-2 was evaluated shortly after the pandemic erupted (8). Specifically, Vero/hSLAM cells were inoculated with SARS-CoV-2 isolate for 2 hrs., followed up with supplementation of 5 μM of IVM. Within 24 hrs. after treatment, there was a 93% reduction of viral RNA in the supernatant and 99.8% reduction of cellular viral RNA, compared to controls. After 48 hrs., there was a further 5000-fold reduction of viral RNA in the supernatant as well as the cell pellets, indicating that cells were essentially ‘cleared’ of SARS-CoV-2 (8). Although IVM possessed a potent antiviral activity (IC₅₀= ~2μM), no cytotoxicity was detected at any time points in this study (8).

IVM was shown to specifically inhibit IMP α/β mediated nuclear import required for replication of HIV-1 and Dengue virus, and therefore it was proposed as the potential mechanism by which it inhibits SARS-CoV-2 (36). Indeed, this baton was passed on, and a subsequent study verified the IVM-IMPα interaction in host cells (38). In fact, it was shown that IVM not only inhibits IMPα association with IMPβ, but can even dissociate IMPα/β heterodimers (38). The specific binding target of IVM was identified, using CD spectroscopy, to be the alpha-helical rich ‘armadillo’ (ARM) domain of IMPα. Moreover, as concentrations of IVM increased, alpha helices in the ARM domain became increasingly destabilized. No changes were detected in the structure of IMPβ. They further verified that this observed effect on IMPα impaired its binding to NLS-containing nsp5 from Dengue Virus (38). As such, preventing N interaction with IMPα, is a likely mechanism that contributes to IVM’s ability to hinder SARS-CoV-2 in vitro replication ( Figure 2 ).

A recent molecular docking study demonstrated that in addition to IMPα, IVM showed high binding affinity to the viral RNA-dependent RNA polymerase (RdRp) complexed with RNA helicase compared to other 10 viral targets included in the analysis (43). However, it was later shown that IVM does not bind to viral RdRp in both Zika virus (Z.V.) and Dengue virus (38). It remains to be identified if IVM may bind to RdRp in coronaviruses.

Other mechanisms of IVM action have also been identified ( Figure 2 ). For example, it previously was shown to suppress the production of Interleukin-6 (IL-6) and Tumor Necrosis Factor alpha (TNFα), two major components of the detrimental cytokine storm induced by SARS-CoV-2 (44). Moreover, a study in Syrian hamsters showed that IVM did not affect SARS-CoV-2 viral load but overall dramatically reduced IL-6/IL-10 ratio and modulated infection outcomes (45). Specifically, hamsters that were inoculated with SARS-CoV-2 were subcutaneously injected with IVM (0.4 mg/kg body weight). IVM reduced severity of clinical symptoms in males, but completely eliminated symptoms in females, which suggests a gender-specific effect of this drug and a factor that should be considered in clinical trials. The gender-specific modulation of IVM on cytokines was also apparent. While females displayed lower levels of cytokines such as IL-6, INFγ and TNFα, males on the other hand developed an enhanced production of INFγ. Notably, viral load in nasal and lung tissues, as well as viral replication rate were not altered in either gender after administration of IVM (45). This is in contrast to the finding that IVM significantly blocks viral replication in vitro and it may be attributed to the much higher dose of IVM that was used (8). However, it is important to note that the dose of IVM that was used on the cells (5 μM) is approximately 50-fold higher than the normal C_max associated with one dose of IVM (200 μg/kg) (46, 47). Therefore, it is important to establish a dose-dependent effect of IVM on viral load and safety in human COVID-19 patients at various doses.

Further, IVM was shown to induce an elevated level of IL-6 and TNFα in onchocerciasis patients, two days after a single dose (150 μg/kg body weight) (48). However, this was attributed to the destruction of the parasite microfilariae, which would usually not be a factor in COVID-19 patients.

Thus far, studies on IVM highlight that it remains important to identify the specific dose of IVM that may reduce viral load, without adverse effects, in humans and to understand if it will differentially affect male and female COVID-19 patients.

The direct toxic effects of IVM were first identified in animal studies, mainly as an antiparasitic treatment. The vast amount of evidence around the use of IVM exists using dose regimens of 150-200 μg/kg of body weight. Hence, the risks and associated side effects are mostly reported at these doses. Studies suggest the common adverse effects are rash, headache, nausea and dizziness, while transient tachycardia is rare and self-limiting (49). Other effects include ataxia, sweating, tremors, and in some cases, coma and death (50).

A retrospective study looking at residents of an extended care facility showed increased rates of death in patients treated with IVM for resistant scabies. These study results were criticized due to some significant limitations of the study, and therefore, the deaths of these residents could not be reliably attributed to the IVM. For example, there was no control of the lasting previous drugs used to treat the scabies, some of which are known for their toxic effects. Importantly, IVM’s toxic effects are short term and are usually resolved (51).

Studies exploring the adverse event profiles of patients on high doses of IVM have also been conducted. Higher dose levels (300-1000 μg/kg) were administered to healthy individuals with head lice, as part of a double-blind and randomized trial, with adverse events reported as having no clinical or biochemical significance (50).

Soon after IVM emerged as a potential therapeutic agent, clinical trials on COVID-19 patients ensued. However, the available published data and ongoing clinical trials, which are summarized in Table 1 , do not provide a clear and uniform understanding of the effect of IVM on COVID-19 patients. This is mainly due to small sample sizes (n=12-203) and the lack of information specifying when exactly IVM is administered after testing positive for SARS-CoV-2 (46, 52–54). It is important to highlight how soon after testing positive the patient receives IVM, in addition to the degree of COVID-19 severity, in order to understand if the effect of the drug is dependent on time and symptom severity. Additionally, several studies are retrospective in which investigators examined past COVID-19 patients who were prescribed IVM, without proper placebo control groups (46, 53). Moreover, most of the studies utilize the antiparasitic effective dose for IVM (0.2 mg/kg body weight), which is substantially less than the equivalent in vitro dose of IVM used against SARS-CoV-2 (8, 53, 54). Nevertheless, the available data does indicate that IVM may, in fact, be effective against COVID-19.

One of the first published studies involved a randomized, controlled double-blind study on 72 hospitalized COVID-19 with mild symptoms (52). Patients that were admitted to the hospital within the last 7 days were either treated with IVM alone (12 mg for 5 days), IVM and doxycycline (12 mg for 1 day, 200 mg doxycycline on day 1 and 100 mg doxycycline every 12 hrs. for days 2-6), or with placebo. The most significant effect of IVM was detected for the rate of viral clearance, measured by a negative rRT-PCR on nasopharyngeal swab. Specifically, the 5-day IVM treatment group demonstrated the fastest rate of viral clearance (approximately 10 days; p<0.02), compared to placebo (approximately 13 days). However, there was no significant difference between the groups for symptoms like cough, sore throat and fever and adverse drug effects (52). Limitations for this study include the exclusion of patients with underlying morbidities and lack of follow-up for mortality and ICU transfers after day 7.

Another recently published clinical study in Florida involved 280 hospital-admitted patients who developed COVID-19 during admission (53). However, it was based on a retrospective analysis of patients and therefore lacked adequate controls. Patients were grouped according to whether or not they received a single dose of IVM (200 μg/kg body weight) and standard care (hydroxychloroquine and azithromycin, unspecified dose) or those who only received standard care. It was not stated, though, at which day post-PCR testing that patients received the drug. The most prominent effect of IVM was the reduction in mortality in the treatment group (15% mortality, p= 0.03) compared to the control group (25% mortality). Mortality was especially reduced in the severe subgroup of patients receiving oxygen support (38.% mortality in IVM group, compared to 80.7% mortality in non-IVM group; p = 0.001). In fact, hospital stay was similar in both groups, as was the rate of viral clearance. However, data was lacking for clinical symptoms as it was not a main outcome (53). In contrast to the previous study, this trial included patients with comorbidities and no adverse drug events were reported. Although this study is limited because it is retrospective, it suggests that IVM may be beneficial in reducing mortality almost by one half, especially for patients receiving oxygen support. Nevertheless, it remains ambiguous as to how soon after testing positive for SARS-CoV-2, patients received IVM, which would have been an important guideline. The results of the few addition published IVM trials are summarized in Table 1 .

There are currently around 50 clinical trials taking place registered on clinicaltrials.gov, which study the effect of IVM as a prophylactic or therapeutic drug (35). Five studies have completed testing at Phase 1, 2 or 3 and three have posted their real-time results, which are included in Table 1 . One of the most promising outcomes include 0% mortality in COVID-19 patients who developed pneumonia (National Clinic Trial Number NCT04343092), however there is no report of an adequate control group and its corresponding rate of mortality. Another ongoing trial (NCT04523831) on COVID-19 hospitalized and non-hospitalized patients also demonstrated 0% mortality, in the IVM group, compared to 1.67% mortality in the control group. Like published studies, ongoing clinical trials also do not present thorough outcomes and the significance statistics are still lacking. As such, more trials are needed which include proper placebo control groups, testing of various doses and records of numerous outcomes.

In addition to IVM, many clinical trials have been conducted to test drugs for COVID-19, many of which have been concluded (47). Most of these medications did not result in significantly improved outcomes, however a few drugs were associated with slightly beneficial effects.

Remdesivir (RMV) has previously been shown to target the viral RNA dependent RNA polymerase in SARS-CoV-1 (55). In the context of SARS-CoV-2, RMV prescribed once daily for 10 days (200 mg day 1, 100 mg days 9-10), showed a survival benefit by days 15 and 28 in patients who did not require oxygen support (56). On the other hand, a large global study by the WHO did not find any clinical benefit for RMV (57). However, an in vitro study suggests a possible synergistic effect of combined RMV and IVM. Specifically, EC₅₀ of approximately 2.3 μM IVM and 1.9 μM RMV were shown to disrupt viral cytopathic activity (58). Because RMV is FDA-approved for the treatment of COVID-19, it is warranted to explore its effect in combination with IVM in clinical trials.

Dexamethasone (6 mg for 10 days) decreased mortality only in severe cases requiring oxygen and mechanical ventilation but was ineffective for mild cases and did not result in any adverse effects (59). Interestingly, a clinical trial (NCT04425863) involving a combined therapy of IVM (0.6 mg/mL solution), dexamethasone (4 mg injection), aspirin (250 mg tablets) and enoxaparin (injection) did indicate a favorable outcome. All patients with mild COVID-19 symptoms (n=135) fully recovered and their symptoms did not worsen. Of those who entered the study displaying severe symptoms (n= 31), one patient perished.

The use of convalescent plasma is also not entirely promising as preliminary analysis based on 1873 reported deaths among 10,406 randomized patients, shows no significant difference in the primary endpoint of 28-day mortality (18% convalescent plasma vs. 18% usual care alone, p=0.34). Although some early studies showed some clinical benefits for convalescent plasma in COVID-19 patients (47), a recent press release from the largest randomized clinical trial, known as the ‘RECOVERY Trial’, revealed otherwise (60). The investigators concluded no evidence of benefit for convalescent plasma in treatment of COVID-19, whereby the 28-day mortality did not differ significantly between the treatment and the control groups. Recently two clinical trials showed that monoclonal antibodies against the spike protein can disrupt progression of early COVID-19 infection (61, 62). However, this type of therapeutic remains very expensive and largely unavailable.

Finally, a recent study, still awaiting peer-review, demonstrated that treatment with the IL-6 receptor antagonists, tocilizumab and sarilumab, improved the clinical outcome including survival in critically ill COVID-19 patients (63). However, these drugs remain expensive and not widely available especially in poor and developing countries.