Anti-pathogenic potential of a classical ayurvedic Triphala formulation

A classical ayurvedic polyherbal formulation namely Triphala was assessed for its anti-pathogenic potential against five different pathogenic bacteria. Virulence of four of them towards the model host Caenorhabditis elegans was attenuated (by 18-45%) owing to pre-treatment with Triphala Formulation (TF) (≤20 µg/ml). TF could also exert significant therapeutic effect on worms already infected with Chromobacterium violaceum (MTCC 2656), Serratia marcescens (MTCC 97) or Staphylococcus aureus (MTCC 737). Prophylactic use of TF allowed worms to score 14-41% better survival in face of subsequent pathogen challenge. Repeated exposure to this formulation induced resistance in S. marcescens, but not in P. aeruginosa. It also exerted a post-extract effect (PEE) on three of the test pathogens. TF was able to modulate production of quorum sensing (QS)-regulated pigments in three of the multidrug-resistant gram-negative test bacteria. Haemolytic activity of S. aureus was heavily inhibited under the influence of this formulation. P. aeruginosa's lysozyme-susceptibility was found to increase by ~25-43% upon TF-pretreatment. These results validate therapeutic potential of one of the most widely used polyherbal ayurvedic formulations called Triphala.


Background
Antibiotic-resistant bacterial infections are among the most serious public-health threats. Since the emergence and spread of antimicrobial resistance (AMR) is shrinking the utility spectrum of conventional bactericidal antibiotics, there is an urgent need for discovery and development of novel anti-virulence formulations. Traditional Medicine (TM) systems like Ayurveda offer several sophisticated formulations for a variety of disease conditions. One such classical ayurvedic formulation with a long history of safe use is Triphala. Triphala is a polyherbal formulation containing three myrobalans fruits i.e. Phyllanthus emblica, Terminalia bellerica and Terminalia chebula (Patwardhan et al., 2015). TF is prescribed as a general health promoter, for management of metabolic disorders, dental and skin problems, and for wound management. It has been reported to be active against bacterial pathogens of urinary tract (Bag et al., 2013), and as an anticaries agent for control of gum infections (Bhattacharjee et al., 2015; Prakash & Shelke, 2014). Though many popular formulations like Triphala have been used historically in TM and as a household remedy, their validation through modern scientific methods is necessary for their wider acceptance in modern medicine (Kothari, 2018). This study aimed to investigate the anti-pathogenic efficacy of TF against five different pathogenic bacteria.

Test formulation
Triphala formulation (TF) (Emami Ltd; batch no. EM0029; Proportion of 3 constituent plant species: 1:1:1 i.e. P. emblica, T. bellerica, and T. chebula) was purchased from a local market in Ahmedabad, India). For assay purpose, 150 mg of this formulation was suspended in 5 ml of DMSO (Merck, Mumbai), followed by vortexing for 15 min. Then it was centrifuged at 8,000 rpm for 30 min at ambient temperature, and resulting supernatant was collected in a sterile 15 ml glass vial (Borosil) and stored under refrigeration till further use. Remaining pellet was subjected to drying in an oven at 70-80°C until the solvent was completely evaporated, followed by weighing of the dried plant material. Subtracting the latter from the initial weight of 150 mg, the concentration of test formulation in supernatant was calculated to be 22.94 mg/ml. This way the whole formulation was found to contain 70% DMSO soluble fraction, which was used for our experiments.

Bacterial strains
a. Three different types of in vivo assays were done as under, employing the methodology described in reference cited in parenthesis following the assay name:

Amendments from Version 1
Following major modifications have been made while revising the manuscript in line with referee's suggestions: 1. Certain sentences have been re-framed to achieve better brevity.
2. Parts of Figure-8 have been re-ordered as per referee's suggestion. Figure-10(c) have been corrected.

Errors in
4. "Pigment Units were calculated to nullify the effect of any change in cell density on pigment production": This sentence has been added to the legend of Figure-8.

5.
As suggested by the referee, we have replaced 'Triphala formulation' with 'TF' at multiple places. Similarly at many places, he has suggested removal of parentheses, that has also been implemented.
6. Legend of Figure-7 has been re-written to eliminate earlier errors of grammar and sentence structuring.
Any further responses from the reviewers can be found at the end of the article REVISED a. Anti-infective assay (Patel et al., 2018a): TF exposedpathogenic bacteria were allowed to infect C. elegans (L3-L4 stage), and their capacity to kill the worm population was compared with their TF-unexposed counterparts, over a period of 5 days.
b. TF as a post-infection therapy (Patel et al., 2019b): Worms already infected with pathogenic bacteria not previously exposed to the test formulation were treated with TF to see whether the test formulation can exert any therapeutic effect on already infected worms. Assay methods remained the same as described in previous section, except that TF was added into assay wells after allowing bacteria either for 6 h or 24 h to establish infection.
c. Prophylactic assay (Patel et al., 2018b): TF-fed worms were challenged with pathogenic bacteria previously not exposed to the test formulation, and their ability to survive in face of pathogen challenge was compared with their TF-unfed counterparts. C. elegans worms maintained on NGM were kept unfed for 24 h prior to being used for experiments. These worms were then fed with TF by mixing required concentration of this formulation (100 µL of DMSO-dissolved Triphala) with 800 µL of M9 medium and placed in a sterile non-treated polystyrene 24-well plate (HiMediaTPG24-1X50NO) containing 10 worms per well. Duration of exposure of worms to TF was kept to 96 h, followed by addition of 100 µL of pathogenic bacterial suspension of OD 764 = 1.50 measured with Agilent Cary 60 UV-Vis spectrophotometer). Appropriate controls i.e. worms previously not exposed to TF, but exposed to pathogenic bacteria; worms exposed neither to TF nor bacteria; and worms exposed to TF, but not to bacterial pathogens, were also included in the experiment. Incubation was carried out at 22°C. Number of dead vs. live worms were counted every day for 5 days by putting the plate with lid under a light microscope 4X objective (Catalyst Biotech CatScope CS-U207T). Straight worms were considered to be dead. Plates were gently tapped to confirm lack of movement in the apparently-dead worms. On the last day of the experiment, when plates could be opened, their death was also confirmed by touching them with a straight wire, wherein no movement was taken as confirmation of death.
Videos of some of the in vivo assays were captured on the fifth day of the experiment, using an inverted microscope (Nikon

Statistical analysis
All the experiments were performed in triplicate, and measurements are reported as mean ± standard deviation (SD) of 3 independent experiments. Statistical significance of the data was evaluated by applying t-test using Microsoft Excel ® 2013. p values ≤0.05 were considered to be statistically significant. Trial version of GraphPad Prism 7 was used to make Kaplan-Meier survival curve for worms.

Results
In vivo experiments Anti-infective assay. When all the five pathogens were pretreated with 0.5-100 µg/ml of TF before being allowed to attack C. elegans, Triphala formulation (TF) was able to attenuate virulence of all test pathogens except S. pyogenes at

TF as a post-infection therapy.
To test the therapeutic efficacy of TF in already-infected worm population, we first allowed different pathogenic bacteria, not previously exposed to TF, to infect C. elegans either for 6 h or 24 h, and then exposed the infected worms to two different concentrations of TF. TF failed to exert any therapeutic effect on worms infected with P. aeruginosa. However, it could exert significant (p≤0.05) therapeutic effect on worms infected with C. violaceum or S. marcescens. Against S. aureus, TF was effective only if the worms were given TF-treatment early (i.e. 6 hour-post infection) [ Figure 4; underlying data (Patel et al., 2019c)]. TF could also not rescue the worms if they already had a mixed infection (S. aureus and P. aeruginosa).
Prophylactic potential of TF. To investigate whether previous feeding with TF can make the worm population tolerate subsequent challenge with pathogenic bacteria, not treated with TF, better; worms were first maintained in a TF-containing M9 buffer for 96 h, and then challenged with different bacterial pathogens. Such TF-fed worms scored 14.50-41.50% better survival in face of pathogen challenge [ Figure 5; underlying data (Patel et al., 2019c)]. However, TF did not confer any prophylactic benefit on the worm population against mix-culture challenge of P. aeruginosa and S. aureus. Since prophylactic activity of any formulation can be said to stem mainly from its effect on the host, we also compared whether TF imparts any extension of longevity on the worm. Worms fed with TF (10-20 µg/ml) registered marginally better survival up to 11 days ( Figure 6).

Repeated exposure of test pathogens to TF.
Since one of the major challenges with even the most potent antimicrobial agents/formulations is development of resistance against them by the pathogen populations, we tested whether repeated exposure of the test pathogens to TF can induce any resistance in them. For this, we subcultured two of the gram-negative test pathogens (P. aeruginosa and S. marcescens) in TF (50 µg/ml)containing broth for 10 subsequent times, and then the 'TF-habituated' cultures thus obtained were tested for their virulence towards the nematode host. Since antibiotic-resistant strains of P. aeruginosa and Entrerobacteriaceae (to which S. marcences belongs) are recognized as serious threats (https://www. cdc.gov/drugresistance/biggest-threats.html; https://www.who. int/medicines/publications/WHO-PPL-Short_Summary_25Feb-ET_NM_WHO.pdf), these two bacteria were used in this assay. Since higher concentration of antimicrobial formulations can be expected to exert higher selection pressure on the susceptible bacterial population, we chose the highest TF concentration (50 µg/ml) beyond which TF does not exert any statistically superior anti-infective effect.
[ Figure 8; underlying data (Patel et al., 2019c)]. It did so with S. marcescens without affecting its growth, which is characteristic of an ideal anti-virulence agent i.e. attenuation of virulence without exerting any selection pressure on the susceptible pathogen. However, TF did exert a growth-inhibitory effect on P. aeruginosa, wherein its IC 50 was observed to be near 50 µg/ml. The quorum modulatory effect of TF on pigment production in P. aeruginosa was observed not to be amenable to be described by a linear dose-response curve. It seems to fall within the realm of hormesis (Calabrese, 2004). For example, production of both pigments was not affected maximally at the highest test concentration. Pyocyanin production was inhibited more at 0.5 µg/ml than at 20 µg/ml. Effect of TF on pyoverdine production followed a linear threshold model within concentration range of 0.5-50 µg/ml, but it took an inverted-U shape over 20-100 µg/ml. Though the exact Figure 8. Effect of Triphala on growth and QS-regulated pigment production in test pathogens. Bacterial cell density was quantified as OD 750. DMSO (0.5%v/v) in the vehicle control did not affect growth and pigment production in any of the test pathogens. Pigment Units were calculated to nullify the effect of any change in cell density on pigment production (A) S. marcescens: OD of prodigiosin was measured at 490 nm, and Prodigiosin Unit was calculated as the ratio OD 490 /OD 750 (an indication of prodigiosin production per unit of growth). Catechin (100 µg/ml) inhibited prodigiosin production by 13.05%*±0.10 without affecting bacterial growth; Ampicillin (5 µg/ml) inhibited growth and prodigiosin production by 8.48%**±0.02 and 40.60%*±0.23, respectively. (B) P. aeruginosa: OD of pyoverdine and pyocyanin was measured at 405 nm and at 490 nm. Pyoverdine Unit and Pyocyanin Unit was calculated as the ratio OD 405 /OD 750 and OD 490 /OD 750 (an indication of pyoverdine and pyocyanin production per unit of growth). Catechin (100 µg/ml) inhibited, pyoverdine and pyocyanin production by 3.85%*± 0.38 and 12.74%*± 2.60 respectively without affecting bacterial growth; Gentamicin (0.5 µg/ml) inhibited, pyoverdine and pyocyanin production by 10.53%*±2.07 and 57.93%***±6.47 without affecting bacterial growth; (C) C. violaceum: OD of violacein was measured at 570 nm, and Violacein Unit was calculated as the ratio OD 570 /OD 750 (an indication of violacein production per unit of growth). Catechin (100 µg/ml) did not affect growth as well as violacein production; Chloramphenicol (0.5 µg/ml) inhibited growth by 40.31**%±0.44 without affecting violacein production. (D) S. aureus: OD of staphyloxanthin was measured at 450 nm, and Staphyloxanthin Unit was calculated as the ratio OD 450 /OD 750 (an indication of staphyloxanthin production per unit of growth). Catechin (100 µg/ml) and vancomycin (0.1 µg/ml) did not affect growth as well as staphyloxanthin pigment production. (E) S. pyogenes: TF or catechin (100 µg/ml) did not affect the growth when measured as OD 655 ; Chloramphenicol (0.5 µg/ml) inhibited growth by 7.56%**±3.46. *p<0.05, **p<0.01, ***p<0.001; TF: Triphala Formulation; QS: Quorum Sensing. mechanism responsible for a non-linear dose-response relationship is not known, it may be the varying magnitude of bacterial adaptive response at different concentrations of the test agent, which generates such non-linear response curves (Lushchak, 2014).
We also tested the effect of TF on two important virulence traits of the bacterial pathogens i.e. haemolytic activity, and biofilm. Though TF could not curb haemolytic activity of any of the gram-negative bacteria, this activity of S. aureus was heavily inhibited under the influence of TF [ Figure 9; underlying data (Patel et al., 2019c)]. While P. aeruginosa biofilm was not affected by TF, TF was able to reduce biofilm formation by S. marcescens, and S. aureus. When TF was applied on pre-formed bacterial biofilms, it seemed to enhance synthesis of the biofilm matrix material (quantified thorough crystal violet assay), and also the metabolic activity (measured in terms of organism's ability to reduce MTT) of the bacterial biofilm [ Figure 10; underlying data (Patel et al., 2019c)]. It may be speculated that TF-treatment induces stress in the bacterial population residing in biofilm form, and this causes the bacteria to mount stress-response. Slow metabolism is a general characteristic of bacterial biofilms (Singh et al., 2017), but TF seems to have forced the biofilms of two of our test bacteria to enhance the rate of their metabolic activity, as well as synthesis/secretion of biofilm matrix components (e.g. polysaccharides, proteins, and extracellular DNA). Enhanced production of exopolysaccharide and e-DNA is believed to occur in stressed bacterial populations (Chang et al., 2007;Zatorska et al., 2018). Sub-inhibitory concentrations of betalactam antibiotics have been reported to induce extracellular DNA release and biofilm formation in some S. aureus strains (Kaplan et al., 2012).
During the host-pathogen interaction, host defense mechanisms play a determinant role in deciding the outcome of this interaction. Since lysozyme is an important component of the innate defense machinery of human immune system against invading Figure 10. Effect of Triphala formulation on biofilm of test pathogens. Crystal violet assay was performed to quantify biofilm formation or eradication, wherein amount of this dye retained by the biofilm was read at 570 nm after extracting it in ethanol. MTT assay was performed to quantify biofilm viability (metabolic activity), wherein change in colour of the MTT dye owing to bacterial metabolism was read at 570 nm. Catechin (100 µg/ml) for all test bacteria, ampicillin (5 µg/ml) for S. marcescens, gentamicin (0.5 µg/ml) for P. aeruginosa, and vancomycin(0. microbes (Herbert et al., 2007), we also studied whether TF can have any effect on susceptibility of the test pathogens to lysozyme. TF-treated cells of S. marcescens and S. aureus were found to suffer marginal (albeit statistically significant; p≤0.05) increase in their susceptibility to lysis by lysozyme. P. aeruginosa's lysozyme-susceptibility was found to increase heavily (by ~25-43%) upon TF-pretreatment [ Figure 11; underlying data (Patel et al., 2019c)].
Most conventional antibiotics suffer from an inherent limitation of not being selectively inhibitory to pathogenic bacteria, and they simultaneously inhibit resident bacterial members of the human microbiome; which may lead to gut dysbiosis (Wipperman et al., 2017). Thus an ideal anti-pathogenic formulation should exert anti-pathogenic effects without inhibiting indigenous members of human microbiome. We tested TF's effect on three such bacteria (Enterococcus faecium, Bifidobacterium bifidum, and Lactobacillus plantarum) which are part of human microbiome, and also used as probiotic strains. Though TF did not exert any prebiotic potential by promoting growth of the probiotic bacteria, it also had no negative effect on them [Extended data: Figure S1 (Patel et al., 2019c)].

Conclusion
This study has found the classical TF to possess significant anti-infective potential against gram-positive and gram-negative pathogenic bacteria. It was also found to be efficacious as a post-infection therapy as well as a prophylactic measure against bacterial infection. TF can be said to possess a broad-spectrum of anti-pathogenic activity, which seems to partly arise from its ability to interfere with bacterial quorum-sensing. Its prophylactic efficacy indicates that it is not only exerting inhibitory effect on the susceptible bacteria, but also beneficial effect on the host worm, and thus can be described as a combination of immunomodulatory and anti-pathogenic activities in one formulation. Exerting such combined efficacy without displaying any negative effect on beneficial members of human microbiome are key attributes for 21 st century antimicrobials (Laxminarayan et al., 2013). Further investigation for elucidating the molecular mechanisms associated with the biological effects of TF are warranted, with special emphasis on its role in combating AMR. Such traditional medicine polyherbal formulations need not necessarily be thought of as replacement of conventional antibiotic treatments, but more realistically as adjunctive therapies boosting our efforts to tackle AMR effectively. This project contains the following extended data:

Data availability
• Video (a).avi (Video of C. elegans challenged with S. marcescens) • Video (b).avi (Video of C. elegans exposed to TF-treated S. marcescens) Data are available under the terms of the Creative Commons Zero "No rights reserved" data waiver (CC0 1.0 Public domain dedication).
why different concentrations of TF used for pathogens. In response to that we would like to state that all pathogens were challenged with identical concentrations of TF in the range of 0.5-100 microgram/mL. However in figures corresponding to these results, we have showed only non-overlapping lines. Data regarding other concentrations has been mentioned in the figure legends. If we include lines pertaining to all concentrations in the graph, then it becomes very fuzzy and difficult to understand.
Though referee's suggestion of merging Figure-2 with Figure-1 is fully logical, we are unable to implement it, as that will cause multiple lines in Figure-1 to overlap making it fuzzy and difficult to understand. For the same reason, we were not able to put lines for all TF concentrations in the graphs of Figure-1, and had to include that data in figure legend.

3.
Referee has suggested for including the details of Post-Extract Effect (PEE) in 'Method' section. However that was not done as the method is not different than that described for 'anti-infective assay' except that in the PEE assay the host worm was challenged with daughter cells of the TF-treated parent cells; and this small piece of information has already been mentioned in the 'Results' section while first mentioning PEE.

4.
The referee has asked for data of PEE against C. violaceum. However PEE assay was done only with the pathogenic bacteria which are considered as serious threats, i.e. whose pathogenic/ antibiotic-resistant strains are listed as priority pathogens by CDC/ WHO.

5.
The referee has raised a query about the difference in survival % of unchallenged/ unfed worms between figure 5 and figure 6. It should be noted that prior to the prophylactic assay, whose results are depicted in Figure-5, worms were kept unfed for 5 days; whereas prior to longevity assay (results depicted in Figure-6) the worms were kept unfed for 3 days. Hence this additional 2-dyas of non-availability of food in the prophylactic assay explains lower survival of worms in 'control' wells in Figure-5.

6.
As suggested by the referee, we have replaced 'Triphala formulation' with 'TF' at multiple places. Similarly at many places, he has suggested removal of parentheses, that has also been implemented.

7.
Figure-8 has been revised as per his suggestion. 8. Figure-10: Referee has raised query about accuracy of the % values. We thank him for his fine observation, as investigating this has made us locate errors in some % values as well as bar heights. All those errors have been corrected, and we are submitting revised figure. 9.
All other minor corrections/ modifications marked by the referee in the PDF file have also been either implemented, OR we have provided appropriate response/ explanation by inserting 'comments' in the word file of revised version.

10.
We hope that the revised version along with clarifications provided by us will be able to earn referee's and editor's approval.