Controversies’ clarification regarding ribavirin efficacy in measles and coronaviruses: Comprehensive therapeutic approach strictly tailored to COVID-19 disease stages

Table 3 Severe acute respiratory syndrome coronavirus, Middle East respiratory syndrome coronavirus and severe acute respiratory syndrome coronavirus-2 in vitro, in vivo (animal), in silico study results focused on ribavirin treatment

Ref.	Isolates	Cell line	Compounds studied	Results (EC50 μg/mL INFs IU/mL, unless stated otherwise)	Researchers’ conclusion	Comment
SARS
Cinatl et al[113], 2003	FFM-1, FFM-2	Vero	6-azauridine	→ 16.8	Glycyrrhizin inhibits SARS-CoV replication; search for therapeutic compounds against SARS will be greatly facilitated by establishing growth of SARS-CoV in human cells	Utilization of a resistant to RBV cell line (Vero E6 cell line, Pitfall 1); however, authors noticed the need for growth in human cell lines
			Pyrazofurin	→ 4.2
			MPA	→ > 50
			RBV	→ > 1000
			Glycyrrhizin	→ 300
Tan et al[114], 2004	Singapore isolate	Vero E6	IFN-β1b	→ 10000	RBV is inactive against SARS-CoV. IFNs exhibit antiviral activity	Utilization of Vero E6 cell line, Pitfall 1
			IFN-αn3	→ 10000
			RBV	→ 10000
			IFN-β1b + RBV	→ No synergistic inhibitory effect
Ströher et al[112], 2004	Tor2, Tor3 Tor7, Tor684	Vero E6	RBV	→ No SARS susceptibility up to concentrations of 2000 μg/mL	RBV alone is unlikely beneficial; combination with IFN-α2b should be evaluated	Utilization of Vero E6 cell line, Pitfall 1
			IFN-α2b	→ Substantial inhibitory effect at concentrations ≥ 1000
Chu et al[87], 2004	HKU-39849	Fetal Rhesus Kidney-4	RBV	→ 50	Cytopathic effect of SARS was inhibited by Lop and RBV	First different cell line used than Vero → positive results
			Lop	→ 4
Saijo et al[115], 2005	HKU-39849Frankfurt-1	Vero E6	Mizoribine	→ IC50 3.5	Mizoribine and RBV possess an inhibitory effect. Discrepancy between studies for RBV could be attributed to the duration of incubation times of the cells in the presence of RBV	First study with positive results for RBV in Vero cell lines. Totally correct conclusion for discrepancies between studies
			RBV	→ IC50 20
Chen et al[116], 2004	10 SARS-CoV isolates	Vero E6, fRhk-4	IFN-α	→ 5000 (fRhk-4), 19.5 (Vero)	Pre-incubation 16 h with IFNs enhanced activity. RBV and Lop were less active in the Vero cell line. IFNs + RBV was the most effective combination	The study includes Pitfall 1, the rest are ok
			IFN-β1α	→ 2000 (fRhk-4), 10.6 (Vero)
			RBV	→ 50-100 (fRhk-4), > 200 (Vero)
			Lop
			IFNs + RBV synergism	→ 2-4 (fRhk-4), 4-8 (Vero)
Morgenstern et al[117], 2005	FFM1, 6109	Vero, CL14, CaCo2, PK-15, HPEK, MA-104,	RBV	→ > 1000 (Vero), → 9.4 (MA-104), → 2.2 (PK-15), → 5.2-8.2 (human cell lines)	IFN-β + RBV combination inhibits SARS-CoV replication in drastically reduced concentrations	The study includes Pitfall 1, the rest are ok
			IFN-β + RBV	→ 10-fold lower RBV concentration,→ 50-2000-fold, lower IFN concentrations
			Synergism	→ Combination index 0.45
Barnard et al[118], 2006	SARS Urbani strain	Vero 76, Vero E6 MA-104 CaCo2, BALB/c mice	RBV	→ 270, EC90 = 560	RBV, like other IMPDH inhibitors, may enhance viral replication in lungs. Four days after cessation of therapy, RBV promoted pro-inflammatory cytokine production, although 3 d after RBV administration inhibited pro-inflammatory cytokine production in mice by significantly reducing IL-1α, IL-5, MCP-1 and GM-CSF. Authors concluded that their data do not support the use of RBV or other IMPDH inhibitors for SARS treatment	The only study with such high RBV’s EC50 in human CaCo2. It is not explained why animals were treated only for 3 d with RBV showing good results and then it was ceased for 4 d, just to confirm the disease worsening with untoward outcomes. Also Pitfall 1 is included in the study
			RBV	→ 1253
			RBV	→ EC90 = 225
			RBV	→ EC90 = 4100
Shah et al[119], 2010	VSV, SeV	BHK21, BSRT7, HeLa, A549, 4T1, HEp2, Vero	RBV	Both viruses have the ability to initiate infection in all cell line tested. RBV effectively inhibited VSV in BSRT7, HeLa and HEp2 cells even at the lowest tested drug concentrations. However, RBV had a surprisingly mild effect on VSV in Vero and A549 cells even when used at 1000 μg/mL concentration with a somewhat intermediate effect in 4T1 cells. A similar pattern of RBV-resistance was shown for SeV in BHK21, Vero and A549, suggesting that cellular rather than virus-specific factors determine the dramatic differences in their response to RBV. RBV treatment even at 1000 μg/mL concentration did not produce any statistically significant decrease in cell viability in any of the tested cell lines. The development of cell-based resistance to RBV treatment via decreased RBV uptake can greatly limit RBV activity. RBV uptake was inhibited in most cell lines at both lower and higher NBMPR concentrations, a specific inhibitor of ENT via ENT1, 2, previously shown to be primarily responsible for RBV import into the cells. Dramatic variations were observed in the long-term accumulation of RBV in different cell types. Importantly, it correlated with the antiviral efficacy of RBV in the tested cell lines. All the three RBV-resistant cell lines, BHK21, A549, and especially Vero showed markedly decreased levels of RBV accumulation suggesting that such differences in the intracellular RBV metabolism may be responsible for the natural cell resistance to antiviral RBV treatment. Act-D an inhibitor of DNA-primed RNA synthesis, was able to revert the antiviral effect of RBV against several RNA viruses, with two proposed mechanisms (1) the stabilization of cellular GTP levels, and (2) inhibition of RTP production. ActD had a clear neutralizing effect on RBV in most cell lines. ActD treatment did not inhibit RBV uptake, demonstrating that the observed reversal of RBV antiviral action was not due to interference of ActD with RBV uptake. The observed resistance of VSV and SeV to RBV in Vero, BHK21, and A549 was not due to the generation of RBV-resistant mutants in these cells. Even when the cells were pretreated with RBV starting 24 h before infection, a little effect of RBV on viral replication in RBV-resistant cells was observed, ruling out any possibility of virus adaptation to RBV. In addition, when VSV was passed by 10 to 15 times in HeLa, BSRT7, and BHK21 cells in the presence of sub-inhibitory RBV concentrations, no viral adaptation to RBV was ever observed. RBV uptake in all tested cell lines after 15min treatment showed that no one of the tested cell lines was defective to RBV uptake. In long-term RBV accumulation in cells after 16 h or 24 h treatment four cell lines sensitive to RBV showed significantly higher levels of RBV accumulation compared to RBV-resistant BHK21, A549, and Vero. Vero cells had a particularly low accumulation which explains the highest resistance to RBV. This long-term accumulation is dependent on the cellular metabolism of RBV. Neutral RBV molecule can be transported freely in and out of a cell via ENTs but once it is phosphorylated, negative charged RMP, RDP, or RTP are trapped inside the cells. A good illustration of the difference between the RBV uptake and its long-term accumulation is RBV hyper-accumulation in erythrocytes resulting in hemolytic anemia in some RBV-treated patients. Similarly to nucleated cells, RBV is transported into erythrocytes via ENTs and converted into RMP, RDP, and RTP. However, unlike nucleated cells, erythrocytes lack the phosphatases needed to hydrolyze RMP/RDP/RTP into RBV. Exogenous guanosine had a clear (almost 100%) neutralizing effect on RBV in BHK21, A549 and Vero cells, which are already highly resistant to RBV. However, very little effect was observed on the RBV activities in RBV-sensitive cells, especially HeLa, 4T1, and HEp-2 cells. Unlike guanosine, ActD was able to effectively neutralize RBV in all tested cell lines. Authors hypothesized that RBV antiviral activity in these cell lines depends not only on the depletion of GTP pool (can be restored by guanosine) but also on the successful 5’-phosphorylation of RBV into RMP/RDP/RTP. At the same time they suggested that RBV acts in RBV-resistant cells types primarily via depletion of GTP pool due to insufficient amounts of phosphorylated RBV molecules in these cells, explaining why the effect of RBV can be completely reversed in these cell lines by guanosine
Smith et al[120], 2013	MHV-A59 SARS-CoV (Urbani strain)	Murine astrocytoma DBT Vero E6	CoVs contain the largest known RNA genome and encode an array of 16 viral replicase proteins, including a 3' to 5' exoribonuclease domain, ExoN, within the non-structural protein 14 Nsp14. The exon is the first identified proofreading enzyme for an RNA virus and functions together with other CoV replicases to perform the crucial role of maintaining CoV replication fidelity. In DBT cells, MHV-ExoN+ viruses were resistant to 10 μM of RBV, while MHV-ExoN- virus titers decreased by~200-fold following treatment with same RBV concentrations, a surprising finding because at least 10-fold higher RBV concentrations are required to inhibit poliovirus and chikongunya viruses. Furthermore, they determined the sensitivity of ExoN + and ExoN-at a low multiplicity of infection. Unexpectedly, multi-cycle replication of ExoN-viruses in the presence of RBV was indistinguishable from single-cycle replication. Using qRT-PCR, researchers determined that ExoN-genomic RNA was dose-dependently reduced by RBV, while ExoN + RNA was unaffected. Extracellular addition of guanosine restored ExoN-titers, even in presence of RBV. These data indicate that the antiviral activity of RBV against MHV-ExoN-viruses is occurring, at least in part, through decreasing viral RNA synthesis and inhibition of IMPDH, while the presence of ExoN activity is capable of preventing RBV inhibition of CoV replication. However, the increased sensitivity of MHV-ExoN-to RBV could result from the impairment of undefined functions of ExoN during replication, particularly during RNA synthesis
MERS
Chan et al[153], 2013	hCo-EMC	MDCK	1280 drugs screened		IFN-β1b and MPA should be considered in the treatment trials of MERS. IMPDH inhibitors inactive in Vero cell line	A combination of IFNs with RBV was not tested. Coexistence of Pitfall 1
			MPA	→ 0.17
			RBV	→ 9.99-41.45
			IFN-α2b	→ 6709.8
			IFN-β1a	→ 480.5
			IFN-β1b	→ 17.6
		Vero	RBV, MPA	→ Inactive
Falzarano et al[155], 2013	hCoV-EMC/2012	Vero, LLC-MK2	RBV	→ 41.45 (Vero), 16.33 (LLC)	Lower sensitivity to RBV for LLC than Vero cells. RBV + IFN-α2b, when combined, the inhibitory concentrations drops to ranges achievable in humans	Coexistence of Pitfall 1. Very significant outcomes for IFN + RBV combination
			IFN-α2b	→ 58.08 (Vero), 13.26 (LLC)
			Combination	8-and 16-fold decrease in the inhibitory concentration as either treatment alone
Falzarano et al[156], 2013	hCoV-EMC/2012	Rhesus macaque	IFN-α2b + RBV	Treated animals showed improved clinical parameters, no dyspnea, little evidence on X-ray. They also showed reduced systemic and local pro-inflammatory markers, significant reduction in viral genome copies in lung tissues and less severe histopathological changes compared to untreated	They suggested IFN-α2b + RBV should be considered for early intervention therapy in MERS. The hedgehog signaling pathway was identified as a putative contributor to decreased lung damage	Very significant results for the early IFN-α2b + RBV administration
Hart et al[154], 2014	Hu/Jordan- N3/2012 (Jordan strain)	Vero E6	MPA	→ IC50 = 2.87	INF-β and MPA or a combination should be considered for MERS-CoV infected patients	The study involves Pitfall 1
			RBV	→ > 250
			IFN-β	→ IC50 = 1.37IFN-β antiviral activity was 16- 41-, 83-, and 117-fold higher than those of IFN-α2b, IFN-γ, IFN-type I and IFN-α2a, respectively
Chan et al[157], 2015	EMC/2012	Common marmosets	MMF	→ A single dose did not improve and might have worsened MERS infection	IFN-β and Lop/r are effective against MERS infection in common marmosets. They concluded that potentially effective combinations that should be evaluated could be RBV and IFN- β1b and/or Lop/r. Although high doses of RBV are limited by side-effects, low-dose RBV combined with IFN- β1b and/or Lop/r may be synergistic	RBV was not tested, but surprisingly the authors discuss its combination with IFN-β1b and/or Lop/r as potentially effective
			Lop/r	→ Improved clinical, radiological, pathological features, and lowered lung and other tissue viral load
			IFN-β1b	→ Less severe disease and lower tissue viral loads
SARS–CoV–2
Computational studies
Kandeel et al[183], 2020	The first available crystal structure of COVID-19 proteins is the main protease, M-pro, and belongs to the translated NSPs together with the papain-like protease Pl-pro		20 drugs	Molecular modelling, virtual screening, docking, sequence comparison statistics and phylogenetics of the COVID-19 M-pro were investigated. Phylogenetic analysis showed a 96.08% identity between COVID-19 and SARS-CoV M-pros, while low identity of 51.61% was detected for COVID-19 and MERS-CoV. In the Schrodinger glide docking module, curcumin was found to be a strong inhibitor of SARS M-pro and the tested compounds’ relative docking scores were calculated compared with the docking score for curcumin. RBV and telbivudine were ranked at the 2nd and 3rd positions respectively, where RBV was shown to form two hydrogen bonds with M-pro. Given the high similarity of SARS and COVID-19 M-pros, RBV as well as telbivudine might be of value in treating COVID-19
Elfiky et al[184], 2020	NSPs such as RdRp (nsp12) are crucial enzyme in the life cycle of the RNA viruses. Docking experiments were performed using the optimized COVID-19 and SARS RdRps		Anti-polymerase drugs against HCV	The active site of RdRp is highly conserved, representing two successive aspartate residues protruding from a beta-turn structure, making them surface accessible through the nucleotide channel (which free nucleotides can pass through). Sofosbuvir and RBV are nucleotide derivatives competing with physiological nucleotides for the RdRp active site, and form 7 and 13 H-bonds respectively. Sofosbuvir, RBV and remdesivir can be used against the nCoV-2019, having promising results. GTP derivatives may be used as specific inhibitors against COVID-19
In vitro studies
Choy et al[187], 2020	BetaCoV/ Hong Kong VM20001061/2020	Vero E6	RBV	→ CPE 500 mmol/L	Remdesivir, Lop, and emetine inhibit SARS-CoV-2 replication. RBV and favipiravir showed no inhibition. Combinational therapy may provide better clinical benefits	Pitfall 1
			Remdesivir	→ CPE 25 μmol/L
			Lop	→ CPE 25 μmol/L
			Favipiravir and others	→ CPE > 100 μmol/L
Wang et al[186], 2020	nCoV-2019 BetaCoV/ Wuhan/ WIV04/2019	Vero E6	RBV	→ 109.5 μmol/L	Remdesivir and chloroquine are highly effective in the control of 2019-nCoV	Pitfall 1
			Favipiravir	→ 61.88 μmol/L
			Nafamostat	→ 22.5 μmol/L
			Nitazoxanide	→ 2.12 μmol/L
			Remdesivir	→ 0.77 μmol/L
			Chloroquine	→ 1.13 μmol/L

ActD: Actin-D; CoV: Coronavirus; COVID-19: Coronavirus disease 2019; ENT: Equilibrative nucleoside transport; GTP: Guanosine triphosphate; IFN: Interferon; IL: Interleukin; IMPDH: Inosine monophosphate dehydrogenase; Lop: Lopinavir; M-pro: Main protease; MERS: Middle East respiratory syndrome; MMF: Mycophenolate mofetil; MPA: Mycophenolic acid; nCoV-2019: Novel coronavirus 2019; NSP: Non-structural protein; RBV: Ribavirin; RdRp: RNA-dependent RNA polymerase; RMP: Ribavirin monophosphate; RTP: RBV triphosphate; SARS: Severe acute respiratory syndrome; SeV: Sendai virus; VSV: Vesicular stomatitis virus.