Ayurveda botanicals in COVID-19 management: An in silico- multitarget approach

Swapnil Borse AYUSH-Center of Excellence, Center for Complementary and Integrative Health, Interdisciplinary School of Health Sciences, Savitribai Phule Pune University, Pune 411 007 Manali Joshi (  manali.joshi@gmail.com ) Bioinformatics Center, Savitribai Phule Pune University, Pune 411 007 Akash Saggam AYUSH-Center of Excellence, Center for Complementary and Integrative Health, Interdisciplinary School of Health Sciences, Savitribai Phule Pune University, Pune 411 007 Vedika Bhat AYUSH-Center of Excellence, Center for Complementary and Integrative Health, Interdisciplinary School of Health Sciences, Savitribai Phule Pune University, Pune 411 007 Safal Walia Bioinformatics Center, Savitribai Phule Pune University, Pune 411 007 Sneha Sagar Department of Pharmaceutical Chemistry, L.J. Institute of Pharmacy, Sarkhej, Ahmedabad 382210 Preeti Chavan-Gautam (  chavanpriti@gmail.com ) AYUSH-Center of Excellence, Center for Complementary and Integrative Health, Interdisciplinary School of Health Sciences, Savitribai Phule Pune University, Pune 411 007 Girish Tillu AYUSH-Center of Excellence, Center for Complementary and Integrative Health, Interdisciplinary School of Health Sciences, Savitribai Phule Pune University, Pune 411 007


Introduction
The, SARS-CoV-2 virus is responsible for causing the ongoing Coronavirus disease (COVID-19) pandemic 1 . Higher infectivity as compared to the SARS-CoV virus reported in 2003 and absence of a de nite cure are the worrisome aspects of SARS-CoV-2. 2 . In the view of rapid spread in short period of time, the number of people infected globally is enormous (~ over 2 million), and this poses a tremendous challenge to healthcare systems.COVID-19 has an age related skewed distribution of morbidity and an overall lethality although the numbers are changing as the disease is progressing 3 . The Centre for Disease Control and Prevention (CDC) reported that, COVID-19 patients with co-morbidities such as chronic lung diseases (e.g. asthma, COPD), hypertension, obesity, Type 2 Diabetes Mellitus (T2DM) are vulnerable to a higher mortality rate 4 .
The SARS-CoV-2 primarily attacks lung alveoli for its replication. The Spike protein of the virus binds to Angiotensin Converting Enzyme-2 (ACE-2) receptors on the surface of type-II pneumocytes of alveolar lining, which are then internalised and +ssRNA is released 5 . With the help of host ribosomal machinery and the RNA-dependent RNA polymerase (RdRp) enzyme SARS-CoV-2 synthesizes its polyproteins and multiplies its +ssRNA. The new copies of SARS-CoV-2 are released into the alveolar sac by destroying the infected pneumocytes. The in ammatory mediators released after pneumocyte damage recruit immune cells at the infected site. Macrophages release in ammatory cytokines into the blood leading to vasodilation of blood vessels increasing capillary permeability of endothelial cells. Neutrophils release reactive oxygen species (ROS) and proteases to destroy viruses which also damage normal pneumocytes and generate cellular debris in alveolar space. Thesein ammatory and immune responses result into alveolar consolidation leading to increased respiratory rate followed by cough. The systemic in ammatory response acts as messengers to hypothalamus to increase body temperature 6 .In some patients,the cytokine response goes out of control leading to excessive collateral damage to organs with a possible progression to death 7,8 .
Presently there is no cure for the disease however, the treatment is symptomatic. In some countries, patients are being treated using existing combinations of antivirals used for other viral infections 9 . Clinical evidence explaining the e ciency of these antivirals against SARS-COV-2 are limited and inde nable 10 . Therefore, for developing speci c therapies as well as for boosting speed and scale of clinical evaluation WHO launched Solidarity clinical trial on April 8, 2020. This includes screening of four study treatments in comparison with standard of care. Based on available experimental data, remdesivir, lopinavir/ritonavir, lopinavir/ritonavir with interferon beta-1α, and chloroquine or hydroxychloroquine are chosen study drugs is required to have safe pharmacotherapy for COVID-19 that can be co-prescribed with WHO solidarity trial drugs and commonly prescribed drugs such as anti-hypertensive, anti-asthmatic and anti-diabetic. Administering plasma of a recovered patient to the critically ill COVID-19 patients seems promising 12 .Currently there is limited evidence on the safety and e cacy of Hydroxychloroquine which is being used in many countries for COVID-19 treatmentunder emergency circumstances 13,14 .There is a great rush to nd vaccines and therapeutics against SARS-CoV-2. Several pharmaceutical companies have announced clinical trials for drug and vaccine candidates. However, it may take a long time to reachthe community.
Traditional medicine systems such as Ayurveda, have a holistic approach of consideringmind-body-physiologyto deal with disease conditions 15 . The Ayurvedic philosophy suggests delivering "a group of phytoconstituents" that holds potential to act on multiple targetsof SARS-CoV-2 and also of immune pathways to give adaptogenic and immunomodulatory effect 16,17,18 .
Therefore, the ideal COVID-19 therapy should show (a) anti-viral properties against SARS-COV-2, (b) be safe for concomitantly administered drugs like anti-hypertensive, anti-diabetic, anti-asthmatics, and drugs those are used in respiratory tract infections (c) should modulate immune system with rejuvenation ability (mainly for cardio-respiratory and nervous system) (d)should show therapeutic adjuvant activity with drugs used in WHO Solidaritytrial and COVID-19 associated comorbidities.
In this work, phytochemicals present in AR, TC and WS were identi ed using chromatographictechnique. A network pharmacology model was used to identify and depict the interactions of bioactives with molecular targets in the immune system to unravel their immunomodulatory role. Further, the phytoconstituents were docked to three molecular targets of SARS-CoV-2 to assess the potential for antiviral activity. Further, predictive tools were used to assess the potential of interactions between phytoconstituents and commonlyprescribeddrugs.In summary, we present a compelling case that Ayurvedic Rasayana botanicals have the potential to be used as anti-SARS-CoV-2 agent having immunomodulatory and therapeutic adjuvant activity for COVID-19management.

Results
1. Extraction of test materials and phytochemical analysis: The percent yield of hydro alcoholic extract of WS, TC, and AR were found to be 10%, 6.6%, and 30% respectively. The percent yield of water extract of WS, TC, and ARwere 12%, 07% and 20% respectively.
Total of 31 phytoconstituents were identi ed from these plant extracts using HPLC and LCMS. The AR extract showed the presence of Asparagamine A, Asparanin A, Isoagatharesinol, Muzanzagenin, Rutin, Shatavarin-I,Shatavarin-IV, Shatavarin-IX,  (Mpro) in the PDB at the time of writing this manuscript. The largest share of the deposition is a series of Mpro crystal structures obtained by fragment screening (Fearon, D., Unpublished). This indicates that the binding site of Mprois druggable. The Mprois known to be functional as ahomodimer and has a heart-like shape. The protein has one active site per monomer. The active site contains the catalytic cysteine-Cys145, thatperforms the proteolysis reaction. The structure, PDB ID: 5R84, solved at a resolution of 1.83Å was chosen for this study. In this structure, the protease is cocrystallized with the fragmentcyclohexyl-N-(3-pyridyl) acetamide (Z31792168) which is seen to bind in theS1-S3pocket of the protease (Figure 1a). The S1 pocket is characterized with the presence of His163, Glu166, S2 with Cys 145 and S3 is known as the aromatic wheel and includes Phe181 and His41. Thus, in our studies weplaced the docking grid over the S1-S3 pocket.
Initially, the co-crystallized ligand was separated and docked to check if Autodock was able to reproduce the binding mode.
The docked pose showed an RMSD of 0.46 Å with thecrystal structure pose and adocking score of -6.2 kcal/mol( Figure 1A and B). A detailed study of interactions revealed that the carbonyl group of Z31792168shows a H-bond with Glu166 in the S1 pocket. The cyclohexylmoiety of the ligand has hydrophobic interactions with Met165, His41, His164, Arg188 andGln189 residue in the S3 pocket. Thepyridine ring scaffold shows hydrophobic interactions with Phe140 and His163 in the S1 pocket.
All these interactions are in line with the crystallographic pose.
Thereafter, 31 phytoconstituents were docked to the binding site. The results are compiled in Table 1.It is a frequent observation that compounds with a docking score better than -6 kcal/mol have a higher probability of being active in vitro and in vivoand hence use this score as a cutoff 24 . In our study we observed that, 18 out of 31 phytoconstituents have a docking score better than -6 kcal/mol.The best docking score -9.9 kcal/mol was observed for Ashwagandhanolide. Many other phytoconstituents like Withacoagin, Withaferin and Withanoneare observed to have docking scores close to the -9 kcal/mol. However, a point to be noted is that some of these phytoconstituents are large molecules and hence the docking scores may appear to be in ated due to size and the proportionately larger number of interactions. Ligand e ciency is a measure of the activity corrected for the ligand size 25  Glu166 and the keto group forms hydrogen bond with His163 in the S1 pocket. The ligand forms three additional hydrogen bonds with Gln189, Thr190 andArg188. Additionally, hydrophobic interactions are observed with amino acids Phe140, Asn142, His164, Met165 and Pro186.
2.2 Several phytoconstituents are also predicted to possess good a nity for theRNA dependent RNA polymerase (RdRp):For the RdRp of SARS-CoV-2 the PDB ID: 6M71 was chosen for this study26. This is a cryoEM structure solved at a resolution of 2.9 Å. Residues 367 to 920 of the structure form the RdRp domain. The authors report that a structural comparison of theSARS-CoV-2RdRp with that of Poliovirus (PDB ID: 3OL6) and HCV indicates that the polymerase domain adopts a conserved structural architecture. The residues Arg553, Lys545 and Arg555 form the NTP entry channel while residues Asp760 and Asp761 coordinate divalent cations that stabilize the phosphate group (the cations are absent from PDB ID: 6M71 as it is an apo structure). We placed our grid over the entire RNA binding site based on the structure of Poliovirus RdRp that is cocrystallised with RNA.
Remdesivir is an ATP analog and is being evaluated as a potential treatment for SARS-CoV-2 as an inhibitor of RdRp based on results obtained for MERS and SARS-CoV 27 .
Remdesivir is a prodrug whose active metabolite is GS-441524 (PubChem CID: 44468216). GS-441524 was docked to the RNA binding site. The docked results ( Figure 2 a and b) indicate that the active metabolite forms hydrogen bonds with Asp760 which is an important active site residue and with Lys621 and Tyr619.The position of the active metabolite is that of the NTP entry channel. The docking score for the metabolite is -4.28 and the ligand e ciency is -0.2. Next, 31 phytoconstituents were docked to the binding site. The results are compiled in Table 1. Withanolide-B, Withacoagin, Withanone, Ashwagandhanolide and Muzanzagenin are predicted to possess docking scores ranging from ~ -9 to 10 kcal/mol. A total of twenty-one phytoconstituents have a docking score better than -6 kcal/mol. Calculation of ligand e ciency reveals that the top binders include Muzanzagenin, WithaolideB, Withacoagin and Magno orine. There are sixteen compounds with ligand e ciency better than the active metabolite. However, it is to be noted that the GS-441524 is a chain terminating nucleotide analog, while these compounds would be pure blockers. The detailed interactions of Muzanzagenin with RdRp ( Figure  interactions between the RBD and PD are mediated by residues Gln24, Lys417, Tyr453, Gln474, Phe486, Gln498, Thr500 and Asn501 of RBD. These residues are divided into three clusters, two clusters at each end and one in the center of the RBD. We placed the docking grid around all of these residues in an attempt to nd molecules that could disrupt the contacts between RBD and PD.
The docking results with 31 phytoconstituents reveal that the compounds bind in various areas of the interface region of RBD.
The top scoring compound is Ashwagandhanolide with a docking score of -10 kcal/mol. Ten phytoconstituents have a docking score better than -6 kcal/mol. Withacoagin  The constructed network represents total 19 bioactives from Rasayana botanicals associated with 306 unique human protein targets. Of these, 53 protein targets were found to be involved in 20 immune pathways referred as immune targets ( Figure   4).The distribution of immune targets amongst Rasayana botanicals has been shown in Venn diagram (Figure 5b The compiled data showed 137, 192, and 109 human protein targets of 9bioactives of AR, and 10 bioactives of TC, and WS each respectively. The immune pathway data retrieved 1104 unique protein targets from 20 different human immune pathways of KEGG database (data not showed). Further data analysis showed the association of AR, TC, and WS bioactives with all immune pathways through several protein targets (Figure5). The bioactives i.e. shatavarins of AR, tinosporides of TC, and withanolides of WS play crucial role in immunomodulation 32,8,20 .According to traditional medicine or Ayurveda principles, the synergistic effect of bioactive combination in an extract strengthens physiological immunity superior than a single molecule 33 . We analysed the data of bioactive-target association and mapped against immune pathways to explore the synergism principle by network pharmacology approach 34 . The analysis showed involvement of AR in Th17 cell differentiation with 24 combinations of 8 bioactives and 6 targets. AR was also found to be involved in IL-17 signaling with 13 combinations of 9 bioactive and targets. Data analysis also showed association of TC with chemokine signalling with 13 combinations of 7bioactives and 5 targets. TC was found associated with few other pathways such as NOD-like receptor signaling, leukocyte trans-endothelial migration, and NK cell mediated cytotoxicity by moderate bioactive-target associations.
The analysis also retrieved multimodal involvement of WS in immunomodulation through various pathways. It is found to regulate chemokine signaling at a greater extent with maximum 66 combinatorial associations of 9 withanolides with 11 immune targets. This is followed by potential of WS in FC-gamma R-mediated phagocytosis and receptor signaling ( Predicting Herb-Drug Interactions: Swiss-ADME, chemoinformatics platform was used to predict pharmacokinetics and drug likeliness potential of the 31 phytoconstituents. The Swiss-ADME pharmacokinetic data which also includes the effect of drug metabolising enzymes (CYPs) and transporters(Pgp) on the 31 phytoconstituentswas used to predict the herb-drug interactions if any. It showed that phytochemicals of AR except, Isoagatharesinolmay not inhibit any of the major CYP isoforms. Few TC phytochemicals may act as inhibitor of CYP1A2, CYP2D6 and CYP3A4 whereas few WS phytochemical may inhibit CYP2C9 (Table 2).
Bioavailability RADAR analysis showed, 9 phytoconstituents of TC are orally bioavailable and lie within pink region of graph (supplementary S1 Appendix analyzed phytochemicals indicate lower skin permeability. These properties need to be considered during phytopharmaceutical development. However, these predictions may not be the same for extracts as they contain multiple phytoconstituents. COVID-19 pathogenesis mainly involves respiratory system and afterwards it leads to multiple organ failure based on patient related factors: sex, age, disease, individualization (PRF: SADI). As discussed earlier, diabetic, obese, asthamatic, geriatric, hypertensive population is more prone for COVID-19 36 . These associated comorbidities of COVID-19 are being continually treated along with anti-SARS-COV-2 therapy.Here, we propose the use of Rasayana botanicals for COVID-19 prophylaxis (both pre and post COVID-19) as well as anti-SARS COV-2 activity. This can lead to the chances of HDIs, which maybe harmful/bene cial/fatal 37 . Therefore, with the help of Swiss-ADME data (  Table).This indicatesthat athigher concentrations (generally more than therapeutic dose), these extracts may produce pharmacokinetic HDI in vivo 38 .Based on the above results, these phytoconstituents or Rasayana botanicals may produce bene cial pharmacokinetic-pharmacodynamic interactions in vivo, with the anti-viral and diseasemodifying drugs that are currently being prescribed for COVID-19.

Discussion
The understanding of pathophysiology of COVID-19 is emerging with increasing prevalence over the globe 39 . Severe infections of SARS-CoV-2 lead to mortality due to severe acute respiratory syndrome accompanied with hypoxia followed by organ failure 6 . A wide variation in the patient population rangingfrom asymptomatic, to mild or moderate cases and severe cases (some showing relapse) is reported. In general, we need to have drugs that are best prophylactic (pre and post COVID- 19), immunomodulatory and adaptogenic in nature along with anti-SARS-CoV-2 action. Using in silico approaches, the present study shows that the selected Rasayanabotanicals may have these all the actions and may be effective for management of COVID-19(Figure06).
In the following sections we discuss in detail how these Rasayana botanicals and their phytoconstituents can be potential drugs for the effective management of COVID-19.
Immune-bolstering activity: SARS-CoV-2 attacks pulmonary pneumocytes in alveolar sac for viral replication 7 . Typically, the immune response to viruses involve cytotoxic T-cells.The viral-infected host cells are recognised through binding of T-cell receptors (TCR) to MHC-I molecules. In case of viruses that escape from TCR recognition, physiological immune surveillance employs NK cells to eliminate virus infected cells. Cytotoxicity triggered by cytokines and other cytotoxic mediatorsinduce infected cells to undergo apoptosis 40 . Ouranalysis showed that, withanolides of WS may augment TCR signallingpathway by modulating NF-kappa-β regulators and theta type protein kinase C (PK-C). Withanolides may also involve in NK cell mediated cytotoxicity by modulating beta and gamma type PK-C.Rutin of AR and 20-hydroxyecdysone of TC may also help in destroying infected cells by regulating cytotoxicity mediator TNF-α. Magno orine and menisperine of TC found to modulate apoptosis through caspase proteins. They may involve in mediating innate immunity by regulating proteolytic cascade of complement system through coagulation factors. Shatavarins of AR were found to modulate antigen processing and presentation through heat shock proteins.Withanolides may interact with integrin protein and interleukins to intensify non-in ammatory intestinal IgA secretion to neutralize virus.It has also been reported that Withanone and Withaferin A may disrupt interactions between viral S-Protein and host receptor angiotensin converting enzyme (ACE) by binding to it 22 . Cytokine storm and in ammation: If the body's defence system or prophylactic treatment fails to control the viral entry and its clearance, the body activates strong in ammatory response leading to cytokine storm. Brie y, the internalization of SARS CoV-2 eventually leads to secretion of proin ammatory cytokines to a large extent 41 . This leads to engagement of immune cells at the infected site. The present data retrieved extensive potential of WS to interfere with chemokine signalling through beta and delta type PK-C and chemokine receptors. Similarly, IL-17 plays crucial role in acute and chronic in ammatory responses. The IL-17 signalling can be mitigated predominantly through heat shock protein by AR, glycogen synthase by TC, and prostaglandin synthase by WS.AR derived saponins and TC extract showed anti-in ammatory property by modulating proin ammatory cytokines along with other in ammatory modulators 42,43 . WS extract also modulated cytokine expressions by inhibiting MAPK/NF-κB pathway 44 .This pathogenic reaction may become more complicated in co-morbidities like diabetes, asthma, hypertension, and obesity where in ammasome is already present 45 . This further activates this in ammatory cascade and COVID-19 progression. It has been reported that the Rasayana botanicals containing phytoconstituents like Withaferin-A have an ability to treat the in ammasome 46,47,19,48 . Increased vascular permeability: The macrophage-mediated in ammatory cytokines cause contraction of endothelial cells of blood vessels which leads to increase in vascular permeability 49  Predicted bene cial Herb-Drug Interactions: COVID-19 is a viral pandemic disease with no speci c cure available so far. Few drugs and drug combinations are still under investigation for their e cacy in managing this fatal infection 79,80 .Current treatment involves combination of previously available antiviral agents 81 . COVID-19 patients having co-morbidities like T2DM, hypertension, asthma, obesity, etc. are also being treated with these drugs along with their ongoing prescriptions 4 . Along with modern therapeutic agents, patients are also being treated with drugs from traditional medicinal system like Ayurveda 82,37 .
Thus, while using such diverse treatment regime, there is need for designing and executing detailed DDI (drug-drug interactions) and HDI (herb-drug interaction) studies for proposing safe, e cacious and bene cial combinations. DDI and HDI can be predicted by PK-PD pathways (with special focus on drug metabolising enzymes and transporters) of particular drug.
Cytochrome P450 system (CYP's) are major enzymes involved in catalyzing biotransformation of administered drugs. CYP1A2, CYP3A4,CYP2C9, and CYP2D6are involved in metabolism of around 80-90% of drugs 83 . Pharmacokinetic pro les of plant extracts considered in this study are not well established 84 . Data fromour in silicoanalysis shows some phytochemicals may or may not inhibit main drug metabolizing enzymes ( Table 2). This inference is also being supported by published in vitro studies on human liver microsomes using whole plant extracts of AR, TC, and WS (S2 Table).IC 50 values for all the extracts are > 100 µg/mL which seems to be a higher concentration than used invivo) 85,38 . Thus we speculate thatin vitro studies on human liver microsomes using whole plant extracts of AR, TC, and WS may not show any inhibition of these main CYP isoforms.However, there is a need to explore in vitro -in vivo HDI for rationalizing theiruse in pharmacotherapeutic management of COVID-19 86 . We also recommend that HDI studies should be planned focusing on drugs used in the management COVID-19and its associated comorbidities.The Supplementary S3 ). This raises the concern for careful pharmacotherapeutic management. On the other hand, it can be predicted that there are very less chances of pharmacokinetic mediated HDI with the studied Rasayana botanicals as no such overlap has been found in our in silico studies as well as published in vitro data.
In this study we explored the potential of Ayurveda based Rasayana drugs in COVID 19 management using in silico approaches and propose a library of phytomolecules with the potential to be developed as phytopharmaceuticals. Appropriate pharmaceutics developability assessment needs to be done if any of these phytomolecules are considered for phytopharmaceutical development. The in-silico pharmacokinetic data also shows that, Muzanzagenin is the main drug-like molecule from AR and it has also showed docking score of > -6.0 kcal/mol against all three protein targets of SARS-CoV-2.
Majority of the key phytoconstituents show good oral bioavailability. All phytochemicals of TC were found to be drug-like substances. In case of WS phytochemicals, except Ashwagandhanolide, Withanoside V and Withanoside IV all were found to be drug-like molecules. Though, Ashwagandhanolide has shown highest docking score for all 3 SARS-CoV-2 protein targets its drug-likeness is an issue of concern and needs further detail investigations. On the other hand, the network pharmacology and docking data showed that there are very high chances of bene cial pharmacodynamic HDI for the effective pharmacotherapeutic management of COVID-19. At the same time, we cannot forget that the drug concentration plays the important role in both in-vitro and in vivo studies, therefore effective and safe clinical extrapolation based on experimentally validated data are warranted.
In conclusion, this study provides leads for clinical application of rasyana botanicals in prophylaxis because of their potential in inhibiting the replication of SARS-CoV-2. These botanicals can also be used as adjunct or mainstream treatment for COVID-19, when the disease is manifested with its symptom. The activities on immune mechanisms provide a sound logic for use of these botanicals in treatment. Some of the phytoconstituents have a possible role in arresting disease progression and preventing organ failure by reducing in ammatory responses. The adjunct use of these botanicals poses another question of herb drug interaction and possible reduction in therapeutic effects offered by modern medicines such as remdesvivir, HCQS and other drugs. However, the data suggests there is no harmful HDI. The compiled targets were ltered for Homo sapiens (human) targets for further data analysis.
Simultaneously, KEGG pathway (https://www.genome.jp/kegg/pathway.html) database was mined to obtain proteins/genes involved in twenty human immune pathways 93 . The universally accepted gene names were collected as mentioned before. The entire data was collected and analysed in Microsoft Excel to identify putative immune-associated protein targets of bioactives of AR, TC, and WS. The network of available data was constructed with eloquent representation using Cytoscape 3.7.2 (https://cytoscape.org/) software 94 .
4. Predicting Herb-Drug Interactions:    Part-(D) shows higher probability for the bene cial pharmacokinetic-pharmacodynamic herb-drug interactions37if co-prescribed with WHO Solidarity trial drugs and drugs for COVID-19 associated comorbidities. It is interesting to note that above mentioned phytoconstituents are predicted to have good docking score, ligand e ciency, oral bioavailability, and drug likeliness. This makes them potential molecules for rapid drug discovery and development based on multi-targeted and reverse pharmacology approach for COVID-19 management. However, these in silicopredictions need in vitro -in vivo validation.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. S1appendixRADARgraphsfor31PhytochemicalsofARTCandWS2.docx S1TableStructureandDockingdetailsofthephytoconstituentsfromASTCandWSweretothebindingsite2.docx S2TableInvitroCYPinhibitionData2.docx S3TablePKdataofWHOSolidaritytraildrugsandCOVID19comorbidities2.docx