Host-targeted niclosamide inhibits C. difficile virulence and prevents disease in mice without disrupting the gut microbiota

Clostridium difficile is the leading cause of nosocomial diarrhea and colitis in the industrialized world. Disruption of the protective gut microbiota by antibiotics enables colonization by multidrug-resistant C. difficile, which secrete up to three different protein toxins that are responsible for the gastrointestinal sequelae. Oral agents that inhibit the damage induced by toxins, without altering the gut microbiota, are urgently needed to prevent primary disease and break the cycle of antibiotic-induced disease recurrence. Here, we show that the anthelmintic drug, niclosamide, inhibits the pathogenesis of all three toxins by targeting a host process required for entry into colonocytes by each toxin. In mice infected with an epidemic strain of C. difficile, expressing all three toxins, niclosamide reduced both primary disease and recurrence, without disrupting the diversity or composition of the gut microbiota. Given its excellent safety profile, niclosamide may address an important unmet need in preventing C. difficile primary and recurrent diseases.


The experiments in
are a really nice addition. It is really important to see how NEN affects the gut microbiota alone. In this experiment there is a vehicle control. Was this used in the Figure 3 infection studies or was water used? Can you clarify this?
6. It appears that the vehicle control changes the diversity at least on Day 3, and in the paper it says NEN did not change the diversity. In Figure 4A Day 3, it does? 7. In the text it needs to clearly spell out if it is a water or vehicle control group. Since these mouse models are very complex, it would help to have this consistent in the text and the figure. Also, the details in the methods section would be helpful with this.
8. It appears that the starting microbiota for most of the groups are different at on Day 0. One limitation using this antibiotic treated mouse model is that it does not change the microbiota in a reproducible way, so it is hard to compare across different cages from different treatments, even the same treatments. Figure 4 shows this nicely in all Day 0 time points, and Figure S6 shows this clearly as there are clearly cage affects. This makes the interpretations in this section much harder to make. If the authors want to show that the NEN, is better then the vanco then why do they not just have one figure comparing the vehicle, NEN alone, and vanco alone across each treatment? Then look at the microbiota and ordination to show that the NEN treatment helps the microbiota recover faster or differently then vanco. This seems to be a major conclusion and is not supported by the data currently. 9. Where is C. difficile in the bar plots? Shouldn't the mice be colonized with it? Please point this out to the reader in the text. Figure S7 is not very clear to the reader.

Minor comments
Since this drug is important for helminths and there are studies looking at helminths and the microbiota, I think a sentence in the conclusions putting this into context would be helpful.
Reviewer #2 (Remarks to the Author): The authors report on the discovery of compounds, which block the toxic activity of C. difficile toxins A (TcdA) and B (TcdB). By screening of compounds, which are already used as drugs they identified niclosamide and 2 other related molecules as substances, which inhibited the cytopathic (cytotoxic) effects of TcdB. Niclosamide is an approved drug for worm diseases, which predominantly acts in the gut, because systemic up-take is minimal. They showed that niclosamide has no effect on the glucosylation of Rac protein or on toxin processing by inbuilt cysteine protease activity. Moreover, they report that niclosamide does not directly interact with TcdB or block toxin binding to cells. In contrast, they propose that niclosamide inhibits the acidification of endosomes, which is a prerequisite for toxin translocation into the cytosol of host cells. Not only cell up-take of TcdB and TcdA are blocked but also of CDT, which acts by a completely different toxin mechanism. They show that niclosamide or its more soluble preparation NEN protects mice towards C. difficileinduced pathogenesis. Moreover, they show that niclosamide does not destroy the normal microbiota as it is known (shown) for antibiotic therapy of C. diff-infection e.g. with vancomycin. The finding of the potent action of niclosamide is of some interest. However, the precise mechanism of action of niclosamide has not been studied and clarified. Moreover, the paper contains several unclear statements.

Reviewers' comments:
Reviewer #1 (Remarks to the Author): In their manuscript entitled "Oral niclosamide inhibits C. difficile virulence and prevents disease pathogenesis in mice without disrupting the gut microbiota" by Tam et al. they use a high throughput approach to screen approved FDA drugs that are able to inhibit C. diff toxins. The team found multiple drugs that were able to inhibit toxin, and rationally selected niclosamide for further studies to systematically define the mechanism of action. They go on to show in a very elegant set of experiments that this drug is able to interact with the host; through inhibition of a pore formation process and is able to neutralize TcdB activity. They also show that this drug blocks the activity of TcdA and CDT toxin, which is a significant finding. This section of the manuscript is well done and very exciting. They then ask if this drug is able to alter the gut microbiota, and conduct a series of mouse models concluding that niclosamide or NEN is able to treat CDI, and also prevent recurrence in a mouse model by restoring the gut microbiota after treatment compared to vancomycin treatment.
Most of the conclusions in this manuscript are supported by the data as far as the toxin MOA studies. However, the mouse model studies paired with the microbiota data in Figures 3 and 4 need to be clarified further. The methods section is missing how the mouse model experiments were done in Figure 3 and 4. This does not diminish the major findings of this paper where the authors found a drug that is able to neutralize all three toxins made by C. difficile. This is novel, significant, and will be well received by the wider scientific community. The approach that the authors took to find this drug is also very innovative. The experiment and statistical approaches look appropriate and robust. Again, a large section is missing from the methods at this time.
We thank you for your comments. Regarding the missing methods, we sincerely apologize for these not being included in the submitted manuscript. They somehow disappeared from the penultimate draft. In the current draft, a complete methods section is included.

Major comments
Most of my comments center around the work done in Figure 3, 4, and the methods section.
1. The title for the paper states that this drug inhibits C. difficile virulence and prevents disease pathogenesis in the mouse model. The authors do not show disease in the mouse model only score diarrhea, weight loss, and death. Death is definitely a metric of severe disease, but since this paper is about a drug that can neutralize toxin, I think looking at toxin or damage from the toxin in vivo would have been a priority. Histopathological changes to the cecum and colon with and without NEN and vanco treatment should be a priority to show that the toxin is neutralized in vivo. We are very sensitive to the fact that terms like: 'infection', 'disease', 'associated disease', 'symptoms' and 'outcomes' are often conflated and confused in the literature when it comes to C. difficile. We have gone through the manuscript and have made changes in the main text and the title to be more explicit and precise with our descriptions. With respect to the particular experiments used to demonstrate the mechanism of action of prevention of disease by NEN, we feel that the toxin neutralization via NEN was more than sufficiently demonstrated in this study by the wealth of data generated showing unequivocally that NEN (i) completely inhibited the pathogenesis of all three toxins through a mechanism that is both well validated and well understood; (ii) had no antibacterial effects on C. difficile itself; (iii) did not disrupt the structure or composition of the microbiota; (iv) showed attenuation of toxin-dependent diarrhea and weight loss; and, (v) showed a dramatic protection from the most severe outcome of CDI (i.e, death), which depends complete on toxin action.

2.
In the prevention mouse model in Figure 3 (challenge after 4hr spore gavage), is NEN in water the control or should this be in DMSO? What is NEN resuspended in and if it is not water, should this be the vehicle control as it is shown in Figure 4?
Thank you for catching this -the control is 5% DMSO, and NEN is given in 5% DMSO. This was added in the figure legend.
3. There are no methods on how the mouse model was done or how diarrhea was scored? I think the authors forgot to add this to the manuscript.
Yes, we did. Apologies. We were citing our previous published methods but agree to include the detail methodologies to help readership. We have added the following to the methods section: : C57BL/6 mice (10 per group) were orally administered 10 5 CFU of C. difficile spores from the UK1 (BI/NAP1/027) strain after receiving antibiotic treatment in the drinking water for 3 days, as shown in Figure 3. The antibiotic mixture contains: kanamycin (0.4 mg/mL), gentamicin (0.035 mg/mL), colistin (850 U/mL), metronidazole (0.215 mg/mL), and vancomycin (0.045 mg/mL) was prepared. This corresponds to the approximate daily dose used for each antibiotic such as kanamycin (40 mg/kg), gentamicin (3.5 mg/kg), colistin (4.2 mg/kg), metronidazole (21.5 mg/kg), and vancomycin (4.5 mg/kg). Mice were given intraperitoneal injection of clindamycin (10 mg/Kg) one day before spores challenge. Mouse weights and the development of disease symptoms were monitored daily. Animals that became moribund or lost >20% of their body weight were euthanized. The mice were divided into the following groups: Control (water with 5% DMSO), NEN (2, 10, or 50 mg/Kg suspended in water with 5% DMSO) Recurrent CDI was performed follow our previous established model (Sun et al, IAI, 2011) with modifications: C57BL/6 mice (10 per group) were prepared for primary CDI model as previously mentioned; however, mice were orally administered vancomycin (0.45 mg/mL, starting from day 1 post spore challenge) in the drinking water for 6 days post spores challenge. Mice were given regular water for until the end of the study. Mouse weights and the development of disease symptoms were monitored daily. Animals that became moribund or lost >20% of their body weight were euthanized. Diarrhea was scored as following: 0 (normal hard fecal pellet); 1 (hard to produce fecal pellet yet no rectal inflammation); 2 (liquid feces, inflamed rectum, soiled tail) 4. The recurrence model that is presented in Figure 3E-H is not really a recurrence model as the authors do not look at disease in the mice before they start treatment with vanco or NEN. Recurrence or relapse occurs when patients have disease from CDI, significant inflammation, and then are treated with vanco. The disease resolves and then it comes back again worse then before. There are better mouse models of recurrence. I think this is a model that is looking NEN in concert with vancomycin. Even with treatment the mice still lose weight at day 2 and 11. It would be helpful to know to know what the C. difficile bacterial load was at this point, and also disease via inflammation in the large intestine. If NEN resolves inflammation in vivo then this is again another significant finding and supports the in vitro work in Fig 1 and 2.
We have monitored disease before and during the treatment with vanco ( Figure 3F and 3H) and as it showed that the disease symptoms (weight loss and diarrhea) were apparent before we started to the treatment of vancomycin, which allowed us to have consistent recurrent CDI. We did not wait to treat mice when the diseases became the most severe (day 2-3 post challenge) since at that time around 50% mice would be moribund and have to excluded from the experiments. The schedule of NEN treatment was used as the same as in the primary CDI model since the dose and schedule gave us a consistent reduction of disease severity. The recurrent CDI model used is meant to simulate the clinical setting where patients receive vancomycin treatment after the disease. The disease may come back when patients stop the antibiotic course and disease severity may vary. Herein, mice are given vancomycin 24hrs post spores challenge and evaluate whether NEN will help to reduce the recurrent disease. During the early treatment (day 2) and shortly after the antibiotic withdraw, the disease symptoms were still evidental in NEN treatment group, but significantly lower than those in control group. Figure 4A are a really nice addition. It is really important to see how NEN affects the gut microbiota alone. In this experiment there is a vehicle control. Was this used in the Figure 3 infection studies or was water used? Can you clarify this? Yes. The vehicle control is water with 5% DMSO throughout the entire study.

The experiments in
6. It appears that the vehicle control changes the diversity at least on Day 3, and in the paper it says NEN did not change the diversity. In Figure 4A Day 3, it does? We agree that gut microbiota diversity in NEN-treated and control mice are slightly but statistically significant different on day 3. However, the difference is small in comparison to the effect of Vancomycin on gut microbiota diversity at day 3. The difference between NEN-treated mice and control mice disappeared on day 6, while the differences between these two groups and vancomycin treated mice remain large. Supplementary Figure 5A (Ordination plot) clearly shows the minimal effect of NEN on the gut microbiota structure. We have clarified the sentence in this section, which was changed from: "As shown in Fig. 4a, NEN treatment did not affect the high diversity, composition, or structure ( Supplementary Fig. 5) of the gut microbiota compared to vehicle control, whereas vancomycin treatment dramatically lowered the diversity of the microbiota, shifting the composition to high relative abundance of Lactobacillaceae and Enterobacteriaceae (Fig. 4a) as seen previously 37 ." to: "As shown in Fig. 4a, gut microbiota diversity in NEN-treated and control mice are only slightly but statistically significant different on day 3, however, the difference is minimal compared to the large reduction observed after vancomycin treatment. This small difference disappears on day 6, while that observed in the vancomycin-treated group remains high. Further, ordination analysis shows the minimal effect of NEN and the large effect of vancomycin on the gut microbiota structure at all time points post-treatment compared to controls (Supplemental Fig. 5A). Vancomycin treatment dramatically lowered the diversity of the microbiota, shifting the composition to high relative abundance of Lactobacillaceae and Enterobacteriaceae (Fig. 4a) as seen previously37. Thus, we conclude that NEN treatment has a minimal effect on the gut microbiota." 7. In the text it needs to clearly spell out if it is a water or vehicle control group. Since these mouse models are very complex, it would help to have this consistent in the text and the figure. Also, the details in the methods section would be helpful with this.
Fair point. This was added to the figure legend in Figure 3. Mouse infection models were added to the Materials & Methods section 8. It appears that the starting microbiota for most of the groups are different at on Day 0. One limitation using this antibiotic treated mouse model is that it does not change the microbiota in a reproducible way, so it is hard to compare across different cages from different treatments, even the same treatments. Figure 4 shows this nicely in all Day 0 time points, and Figure S6 shows this clearly as there are clearly cage affects. This makes the interpretations in this section much harder to make. If the authors want to show that the NEN, is better then the vanco then why do they not just have one figure comparing the vehicle, NEN alone, and vanco alone across each treatment? Then look at the microbiota and ordination to show that the NEN treatment helps the microbiota recover faster or differently then vanco. This seems to be a major conclusion and is not supported by the data currently.
The reviewer correctly points out the limitation of the model, which is characterized by a high variation on day 0 between experiment due to cage effect (see Supplemental Fig. 6). It is true that this is an inarguable inherent feature of this antibiotic treated mice model system. However, this model system is currently the most applicable model system in translational microbiome research and was recommended for studies that aim to investigate the effect of disruption on the gut microbiota (Gut Microbes. 2016; 7(1): 68-74. doi: 10.1080/19490976.2015.1127463). That said, the conclusions drawn from each independent experiment remain valid as within experiment results are consistent. For example, the recovery of the gut microbiota diversity in NEN treated mice is observed after the antibiotic pre-treatment and in the absence of vancomycin treatment and that independently from the composition of the gut microbiota on Day 0. Conversely, vancomycin consistently, reduces gut microbiota diversity which does not recover on day 6. These findings are supported by ordination analysis as shown on Supplemental Figure 6. Thus, our conclusion that NEN treatment contributes to the recovery of the gut microbiota compared to vancomycin treatment, which does not, remains correct. 9. Where is C. difficile in the bar plots? Shouldn't the mice be colonized with it? Please point this out to the reader in the text.
For clarity reasons, we have only displayed the family-level taxonomic group that were present at 1% or above. C. difficile belongs to the family Peptostreptococcaceae which is detected in proportion >1%, thus are included in the "other" category. We did not expect to detect C. difficile in feces on Day 0, as sampling is performed right after infection in all the experimental setups. Further, in all treated mice with NEN or vancomycin, it is unlikely that C. difficile survived the treatments and thus it would not be detected on days 3 or 6. In untreated controls, the mice do clear the infection which would limit our ability to detect C. difficile. It is important to note that we do detect C. difficile but in abundance that are less than 1%. We have modified part of the legend of Fig. 4 which now reads: "… Each barplot indicates the mean relative abundance of bacterial families with relative abundance >1% from mice in two experimental cages. The Peptostreptococcaceae family of bacteria includes C. difficile and its abundance was consistently below 1% in all experiment and thus is included in "others". …" Minor comments Figure S7 is not very clear to the reader.
We have included a detailed figure legend, which was missing, to help describe this figure.
Since this drug is important for helminths and there are studies looking at helminths and the microbiota, I think a sentence in the conclusions putting this into context would be helpful. Excellent suggestion. We have added a sentence to the end of the paper.
The authors report on the discovery of compounds, which block the toxic activity of C. difficile toxins A (TcdA) and B (TcdB). By screening of compounds, which are already used as drugs they identified niclosamide and 2 other related molecules as substances, which inhibited the cytopathic (cytotoxic) effects of TcdB. Niclosamide is an approved drug for worm diseases, which predominantly acts in the gut, because systemic up-take is minimal. They showed that niclosamide has no effect on the glucosylation of Rac protein or on toxin processing by inbuilt cysteine protease activity. Moreover, they report that niclosamide does not directly interact with TcdB or block toxin binding to cells. In contrast, they propose that niclosamide inhibits the acidification of endosomes, which is a prerequisite for toxin translocation into the cytosol of host cells. Not only cell up-take of TcdB and TcdA are blocked but also of CDT, which acts by a completely different toxin mechanism. They show that niclosamide or its more soluble preparation NEN protects mice towards C. difficile-induced pathogenesis. Moreover, they show that niclosamide does not destroy the normal microbiota as it is known (shown) for antibiotic therapy of C. diff-infection e.g. with vancomycin. The finding of the potent action of niclosamide is of some interest. However, the precise mechanism of action of niclosamide has not been studied and clarified. Moreover, the paper contains several unclear statements. Specific comments 1. I miss time curves of the intoxication of cells. In some experiments it seems that niclosamide blocks completely the effect of TcdB (e.g., Fig. 1f and g). However, one need to have time courses of the intoxication process to see this in detail. This is a good point. To address the time course of intoxication and inhibition in greater detail, we performed a kinetic TEER assay with a fixed dose of TcdB (5pM) and varying NEN concentrations and show the results in the new Supplementary Figure 2c. These data illustrate the kinetics of intoxication showing that after 5h there is complete loss of TEER and that NEN dose-dependently restores TEER at 7h.
2. Intoxication of cells was performed in serum free medium. This enhances generally the toxin activity. However, serum might completely block the effect of niclosamide. Therefore, also these conditions are important. Indeed, this is very important and something that we considered carefully when designing these experiments. For the intoxication data presented in the manuscript, we chose to omit serum from the in vitro assay to better reflect conditions in the gut in vivo since there would not be expected to be any serum, or serum albumin for that matter, in the lumen of the gut. Interestingly, it has been speculated that protein binding that results once niclosamide and NEN enter the blood could contribute to the safety and tolerability in humans. As shown below, both niclosamide and FBS are less potent in the presence of FBS. These data have been added to the Supplementary Information with the creation of a new panel (d).