Indian researchers from the city of Bangalore (Bengaluru) in Karnataka-India have found in a new study that niacinamide or Vitamin B3 (niacin) can be used to enhance a natural occurring antiviral in humans called cathelicidin (LL37) to destroy the membrane of the SARS-CoV-2 virus, thus inhibiting its replication.
The study findings were published on a preprint server and are currently being peer reviewed. https://www.biorxiv.org/content/10.1101/2022.08.26.505399v1
A case in point is the skin that expresses Ace2 in the epidermis in-vivo5, and keratinocytes in-vitro (Fig S1A) but nevertheless is not considered as a primary route of infection6. In support of this, exposure of human epidermal keratinocytes to SARS-CoV-2 does not result in a productive infection (Fig S1B).
Thus, although epidermal keratinocytes are competent for SARS-CoV-2 infection, they may possess an endogenous defence mechanism to inhibit viral infection. The skin possesses a basal defence mechanism endowed by the constitutive secretion of antimicrobial peptides (AMPs)7.
In particular, the human AMP cathelicidin (LL37) has been shown to target various classes of viruses8,9, including respiratory viruses10. Skin keratinocytes exhibit a higher basal level of LL37 secretion than lung epithelial cells (Fig S1C), which could be one of the factors leading to their lower infectivity by SARS-CoV-2 (Fig S1B).
To ascertain whether LL37 is effective against SARS-CoV-2, we incubated the virus with this AMP and assessed its ability to infect an intestinal epithelial cell (Caco2) as a reporter. We observed a dose-dependent decrease in viral gene expression upon treatment with LL37 (Fig 1A).
This LL37 mediated effect was also observed in other SARS-CoV-2 variants (alpha, kappa, delta, and omicron) (Fig 1B). These observations were confirmed by tissue-culture infectious dose (TCID50) assays (Fig S1D). Previous work has suggested that LL37 can interact with the spike protein and the ACE2 receptor and possibly occlude the interaction surface between them
11. However, LL37 has also been shown to interact with, and aggregate on membranes12. Hence, we hypothesized that LL37 may inhibit viral infection in a Spike/Ace2 independent manner. We thus compared the neutralising capacity of LL37 against viruses of different tropism, namely VSV-G and S1 pseudotyped lentivirus particles. We observed comparable reduction in transduction when both pseudotyped lentivirus particles were treated with increasing amounts of LL37 (Fig S1E).
pathogens13. Enveloped viruses such as coronavirus that assemble virions by budding off from the endoplasmic reticulum membrane have a negatively charged membrane due to a higher content of phosphatidylserine (PS)14,15. To mimic the membrane composition of a generic coronavirus, we prepared three different vesicles in which PS composition was varied according to the published range of ER-derived virions15 (Fig S1F). We observed that increasing the percentage of PS resulted in an increase in the negative surface charge on the vesicles (Fig S1F), which was neutralized by the presence of LL37 (Fig S1G).
These results suggest that the positively charged peptide can coat the outer leaflet of the bilayer by electrostatic interactions. To determine the consequence of the interaction of LL37 with the vesicles, we assayed whether membrane integrity was compromised. Using a fluorescence resonance energy transfer (FRET) based membrane disruption assay (schematically shown in Fig S1H), we observed reduction in FRET (fluorescence recovery at 530nm) when vesicles were treated with LL37 (Fig. 1C).
These results indicate that LL37 is more effective in interacting with and disrupting membranes with a higher negative charge (Fig S1F). Previous reports have also indicated that disruption of vesicle membranes by positively charged polymers leads to vesicle clumping16. We therefore investigated whether disruption of the pseudoviruses and SARS-CoV-2 by LL37 would result in their aggregation leading to increase in particle size as measured by dynamic light scattering (DLS). Consistent with the reported effect of cationic polymers on negatively charged membranes, we observed an increase in particle size of SARS-CoV-2 (Fig 1D) as well as pseudotyped virus (VSV-G and Spike) (Fig S1I) upon treatment with LL37.
Disease severity of several viral respiratory infections has been inversely correlated with LL37 levels 17. Since it has been shown that salivary burden of SARS-CoV-2 correlates with disease severity in patients 18, we compared the levels of secreted LL37 in the saliva of SARS-CoV-2 infected and uninfected individuals (patient information in (Fig S1J). Symptomatic individuals had on average ~3-fold less LL37 than uninfected individuals (Fig 1E). Interestingly, asymptomatic positive patients had equivalent levels of LL37 as uninfected individuals. These results suggest that lower LL37 levels may potentially render individuals more susceptible to a symptomatic infection.
Figure 1. Antiviral activity of LL37 against SARS-CoV-2 variants (A) Effect of increasing LL37 concentrations on SARS-CoV-2 neutralization (qRT-PCR) (n=3) (B) Effect of LL37 on the infectivity
of various SARS-CoV-2 variants (qRT-PCR) (n=3) (C) Membrane disruption assay using virus like vesicles, mimicking viral membrane and PS concentration (FRET) (n=3) (D) Particle size analysis of SARS-CoV-2 in the presence of LL37 (DLS) (n=3) (E) Measurement of LL37 in various patient cohorts (ELISA) (A total of 41 individuals were subdivided into negative (16), positive symptomatic (13), and positive asymptomatic (12) cohorts) [Statistical analysis was done using one-way ANOVA (A,D), two-way ANOVA (B,C), and Kruskal Wallis test (non-parametric ANOVA) (E), *p≤0.05, **p≤0.001, ***p≤0.0001]
Since lower levels of LL37 are associated with the symptomatic COVID-19 patient group (Fig 1E), we speculated that increasing the level of LL37 or enhancing the activity of the existing LL37 might serve as a potential means of combating SARS- CoV-2 infection. One method of enhancing the activity of LL37 is to decrease its inherent self-aggregation12 and thereby increase its bioavailability. A common approach to prevent aggregation is through the use of a hydrotrope such as niacinamide (vitamin B3). It is a generally regarded as safe (GRAS) substance used to increase the solubility, and therefore the activity, of various drugs19. Indeed, we observed that LL37 supplemented with niacinamide exhibited an enhanced potency against infection by different variants of SARS-CoV-2 (Fig 2A, S2A).
To understand the mechanism of LL37 interaction with lipid membranes in the presence of niacinamide, we used atomistic molecular dynamics (MD) simulations. Our simulations support the hydrotropic solubilisation of LL37 by an aqueous solution of niacinamide. We found that niacinamide transiently associated with mainly the hydrophobic and non-polar residues of LL37 (Fig S2B) by both aromatic- π and van der Waals interactions. These predominantly included phenylalanine (Phe5, Phe6, Phe17 and Phe27) and isoleucine (Ile20 and Ile24) residues (Fig 2B), which mediates LL-37’s self-aggregation and reduced activity20. Our simulations suggest that encapsulating of aggregation prone residues of LL37 by niacinamide would likely improve the bioavailability of the peptide.
In addition, we employed molecular dynamics simulation to investigate the early steps of LL37 adsorption on viral envelope-like membranes in the presence of niacinamide. These simulations demonstrate the lipid acyl chains and headgroups interacting with the LL37 peptide (Fig 2C top image). In accordance with previous literature21, these interactions pull the peptide into the membrane resulting in local thinning in the bilayer and ultimately a dramatic destabilization of membrane (Fig
2C bottom, Fig S2C), with a concomitant destabilization of lipid ordering (Fig S2D). This configuration meant that the charged amino acid residues of LL37 faced the solvent and were free to interact with niacinamide (S2E). Additionally, this simulation predicts that niacinamide also penetrates into the membrane (Fig 2D) and may synergize with LL37 to disrupt the membrane. Thus, we conclude that niacinamide has dual activities: (i) hydrotropically increase the aqueous solubility of LL37, thereby rendering it more bioavailable and; (ii) cooperate with the peptide to destabilize membrane.
To validate these computational results, we performed a FRET-based membrane disruption assay using artificial viral membranes (as described in Fig S1F, H) in the presence of LL37 and niacinamide. Consistent with simulation predictions, we observed that membrane disruption of liposomes by LL37 was enhanced in combination with niacinamide (Fig 2E), while niacinamide by itself was not antiviral. To test whether the effect of niacinamide can be reproduced with naturally produced AMPs, we analyzed its effect on AMPs that are highly secreted in saliva22 and from the skin7.
We found that human saliva exhibits antiviral activity against SARS-CoV-2, which can be potentiated upon supplementation with niacinamide (Fig 2F). Likewise, we also observed that skin scrubs supplemented with niacinamide exhibited antiviral activity (Fig. S2C).
Our body naturally synthesizes niacinamide, but interestingly, the biosynthetic pathways and precursor leading to niacinamide production are downregulated in symptomatic COVID-19 patients23. Thus, exogenous supplementation of niacinamide in symptomatic patients may potentiate the activity of naturally produced AMPs from the body’s epithelia.
Figure 2. Effect of niacinamide on the antiviral activity of LL37 (A) Viral gene expression at different concentrations of LL37 in the presence of niacinamide (qRT-PCR) (n=3) (B) Contribution of each residue of LL37 towards interaction with niacianamide from MD simulations (mean of 3 independent simulations, errors propagated from each replicate). Residues contributing > 1 kJ/mol are labelled (C) The LL37 peptide gets deeply embedded in the membrane in 200 ns (top). Representative membrane thickness averaged over the last 20 ns of MD simulations in absence and presence of the LL37 peptide (bottom left and right respectively). Color scale indicates membrane thickness in nm. (D) Representative snapshot (left) of niacinamide penetrating deeper than water into the hydrophobic core of the membrane (LL37, black peptide; lipid acyl chains, grey lines; niacinamide, space filling; water, orange) which is quantified on the right, averaging the molecular density (membrane, black; niacinamide, green; water, orange) over the entire 200 ns trajectory (5 replicates, mean ± SD) (E) Membrane disruption assay on addition of LL37 and niacinamide (FRET) (n=3) (F) Effect of niacinamide addition to saliva on SARS-CoV-2 neutralisation (TCID50) (n=4) [Statistical analysis was done using student t-test for (A), two-way ANOVA for (E), and one-way ANOVA for (F), *p≤0.05, **p≤0.001, ***p≤0.0001]
The variants that exhibit increased transmissibility and disease severity are reported to contain mutations in the receptor binding domain (RBD) of the spike protein24. Given their key role in mediating viral entry into the host cell, many vaccines have been developed with antigens derived from the spike protein, and mutations pose a serious problem with vaccine escape25. In addition, the heavy glycan coating of the spike proteins can be a mechanism of camouflaging them from the immune system26. One approach to circumvent these problems is to target the viral envelope that originates from the host cell and is thus conserved among the different variants. Altogether, we show that the AMP LL37 has the potential to neutralise the SARS-CoV-2 viral infection by targeting its envelope and niacinamide further enhances this antiviral activity of the peptide. Our data on the symptomatic patient samples further substantiate this hypothesis and argues for an approach that would entail enhancing the efficacy of antimicrobial peptides for protection against viral infections. Therefore, either exogenous administration of the AMP with niacinamide or other strategies to boost the endogenous production of the peptide in combination with niacinamide could be a potent method to not only block viral transmission, but may be an effective therapy to limit viral load and disease severity of a patient post infection.
