Valproic acid use is associated with diminished risk of contracting COVID-19, and diminished disease severity: Epidemiologic and in vitro analysis reveal mechanistic insights

Abstract

The SARS-CoV-2 pandemic has caused unprecedented worldwide infections from persistent mutant variants with various degrees of infectivity and virulence. The elusiveness of a highly penetrant, worldwide vaccination strategy suggests that the complete eradication of SARS-CoV-2 is unlikely. Even with the advent of new antiviral agents, the disease burden worldwide continues to exceed current preventative and therapeutic strategies. Greater interest has been placed towards the development of affordable,broadly effective antiviral therapeutics. Here, we report that the small branched-chain fatty acid Valproic acid (VPA), approved for maintenance of seizure and bipolar disorder, has a novel anti- coronavirus activity that can be augmented with the addition of a long-chain, polyunsaturated omega-3 fatty acid, Docosahexaenoic acid (DHA). An EMR-based epidemiological study of patients tested for COVID-19 demonstrated a correlation exists between a reduced infection rate in patients treated withVPA of up to 25%, as well as a decreased risk of emergency room visits, hospitalization, ICU admission,and use of mechanical ventilation. In vitro studies have demonstrated that VPA modifies gene expression in MRC5 cells. Interestingly, VPA correlates with the inhibition of several SARS-CoV2 interacting genes and the greater inhibition of alpha-coronavirus HCoV-229E (a "common cold" virus) and SARS-CoV2. The VPA-DHA combination activates pre-existing intracellular antiviral mechanisms normally repressed by coronaviruses. Gene expression profiles demonstrate subtle differences in overall gene expression between VPA-treated and VPA-DHA-treated cells. HCoV-229E infection caused an intensely different response with a marked induction of multiple intracellular inflammatory genes. Changes in gene expression took at least 24 hours to manifest and most likely why prior drug screens failed to identify any antiviral VPA activity despite in silico predictions. This report demonstrates an interaction between HDAC inhibition and the potent activation of cellular antiviral responses. A foundation now exists for a low-cost, highly effective antiviral strategy when supplemented with DHA.

Fig 1. Viral inhibition assays around the IC50 of VPA by direct inhibition of HDAC2.

(A.)Antibody-based assay infected with SARS-CoV2. (B.) Firefly luciferase assay infected with SARS-CoV2 (C.) Oligonucleotide hybridization-based assay infected with HCoV-229E. (D.) Oligonucleotide hybridization-based assay infected with HCoV-229E with/without VPA preincubation.

Fig 2. Host cellular genes interactions with the SARS-CoV2 proteins.

(A.): Volcano plot for gene expression levels of selected 300 Gordon gene setat 24 hours (left) and 48 hours (right). (B.): Virus-host protein interaction map of the Gordon gene sets that met significance criteria for Viral Assembly, replication and pathogenicity. (C.) Protein levels measured by western blot for selected gene sets PCNT, DNMT1, BRD2, and HMOX1 at 24 hours (top) and 48 hours (bottom). (D.) Differentially expressed genes upon VPA treatment for 24, 48, 72 and 96 hours.

Fig 3. Effect of different polyunsaturated fatty acids (PUFAs) in the inhibition of HCoV-229Eviral replication in MRC5 cells using antibody-based assay demonstrates significant inhibition of viral replication.

(A): Determination of the IC50value of linoleic acid (LA), demonstrated significant inhibition of viral replication. (B): Determination of the IC50value of docosahexaenoic acid(DHA), demonstrated significant inhibition of viral replication. (C): Determination of the IC50value of eicosapentaenoicacid(EPA), demonstrated significant inhibition of viral replication. (D): Determination of the IC50value of alpha-linolenic acid(ALA), demonstrated least inhibition.

Fig 4. Viral inhibition assay in combination of VPA with or without fixed dose (25 μM)of both DHA and LA.

(A): Determination of the IC50value of VPA with or without linoleic acid (LA). (B): Determination of the IC50value of VPA with or without docosahexaenoic acid(DHA). (C): Determination of the IC50value of TrichostatinA(TSA) with or without docosahexaenoic acid(DHA). (D): Determination of the IC50value of Depsipeptide(Dep) with or without docosahexaenoic acid(DHA).

Fig 5

(A) Percentage of viral RNA sequence detected in RNA-seq from total RNA content of MRC5 cells infected with HCoV-229E at different experimental conditions and different drug combinations. (B) Number of differentially expressed genes (DEGs) identified in MRC5 cells when treated with different drug combinations. MRC5 cells, drug-only compared to mock-treated (Left). MRC5 cells infected with HCoV-229E treated with drug compared to MRC5 cells infected with HCoV-229E alone (Right). (C) K-means clustering scatterplot comparing the Log2(fold change) expression for each treatment group (VPA, DHA, VPA+DHA) versus the MRC5 control cells. (D) K-means clustering scatterplot comparing the Log2(fold change) expression for each treatment group (VPA, DHA, VPA+DHA) versus HCoV-229Einfected MRC5 cells. (E) Principle Component Analysis (PCA) of the differential gene expression obtained underdifferent treatment conditions.

Fig 6. Scatter plots of the molecular pathways obtained from ingenuity pathway analysis (IPA) under different treatment combinations of DHA and VPA.

MRC5 cells were treated with (A) 25 μM DHA, (B) 25 μM DHA and infected with HCoV-229E, (C) 0.5 mM VPA, (D) 0.5 mM VPA and infected with HCoV-229E, (E) 25 μM DHA + 0.5 mM VPA, (F) 25 μM DHA + 0.5 mM VPA and infected with HCoV-229E, (G) 25 μM DHA + 0.5 mM VPA and preincubated with HCoV-229E, (H) MRC5 Infected with HCoV-229E (no drug).

Fig 7

(A.) Percentage of viral RNA detected by qRT-PCR in MRC5 infected with SARS-CoV2 when treated with VPA, DHA, and VPA+DHA. (B.)Percentage of normalized viral RNA detected by RNASeqfrom the cell lysates treated with VPA, DHA and VPA+DHA.