Botulinum neurotoxins (BoNT) are the causative agents of botulism disease, which can be lethal mainly due to respiratory failure (Johnson, 2019). Their extreme toxicities and ease of production led to the classification of BoNTs as Category A bioterror agents, and BoNTs have been attempted for use as biological weapons in the past (Arnon et al., 2001). On the other hand, formulations of these toxins are extensively utilized in clinics as FDA-approved therapeutics to treat various conditions such as movement disorders as well as for cosmetic purposes (Choudhury et al., 2021). BoNT intoxication primarily leads to inhibition of the motor neuron-muscle connectivity by cleaving specific SNARE proteins crucial for neuroexocytosis (Montal, 2010). After these toxins have gained entry into motor neurons, there are no treatment options to inhibit their enzymatic activity inside the cell (Lin et al., 2019). There are at least seven different BoNT serotypes with an increasing number of subtypes and chimeric toxins (Dong and Stenmark, 2021), of which serotypes A, B, E, and F lead to human botulism (Maslanka, 2014). Among these, serotype A (BoNT/A) causes the most human botulism cases, and it is the most utilized BoNT serotype in clinical formulations (Choudhury et al., 2021). This serotype has the most prolonged half-life in neurons compared to other serotypes (Shoemaker and Oyler, 2013). For example, both BoNT/A and another human botulism-causing serotype, BoNT/E, target the same protein, SNAP-25, in neurons; however, BoNT/E is destroyed naturally within a few days to weeks while BoNT/A can be active for up to 6 months in the neuronal cytosol (Shoemaker and Oyler, 2013).

Recent work elucidated the molecular mechanisms by which BoNT/A Light Chain (A/LC) evades the ubiquitin-proteasome pathway (UPP) in neurons (Tsai et al., 2017). More specifically, A/LC is ubiquitinated by the E3 ligase, HECTD2, but it is protected from proteasomal destruction by the dominant effect of VCIP135/VCPIP1, a deubiquitinase (DUB) (Tsai et al., 2017). Notably, previous work provided proof-of-concept that A/LC turnover in cells can be modified by increasing its ubiquitination via designer E3 ligases (Tsai et al., 2010) or by delivering fusion proteins, including a single-chain antibody specific to A/LC fused to an F-box domain recognized by an endogenous E3-ligase (Kuo et al., 2011). However, to pave the road for drug development studies, it is crucial to identify small molecules that can modulate UPP to affect A/LC toxicity. There have been extensive drug development efforts against BoNT/A, and although elegant work from many groups has identified various compound classes as BoNT/A inhibitors, none has graduated to clinical trials yet (Lin et al., 2019). To the best of our knowledge, targeting UPP with small molecules to modulate the persistence of A/LC has not been explored. Although UPP is considered a highly challenging drug target, there are already FDA-approved and many promising clinical/preclinical stage compounds targeting different components of UPP (Wu et al., 2020). For example, proteasome inhibitors Bortezomib, Carfilzomib, Ixazomib, and E3 modulators Thalidomide, Lenalidomide, and Pomalidomide are FDA approved (Wu et al., 2020). In the UPP pathway, DUBs are considered attractive drug targets with significant clinical potential as it may be possible to develop selective DUB modulators due to their diversity (about 100 DUBs in humans) (Basar et al., 2021), and well-defined catalytic clefts (Clague et al., 2019). Indeed, selective DUB modulators, although a few, have been developed (Harrigan et al., 2018; Clancy et al., 2021).

In this work, we focused on identifying small molecules to potentially affect ubiquitination and/or deubiquitination of the A/LC to modulate the A/LC activity in cells. Previous work demonstrated that, in addition to VCIP135/VCPIP1, another DUB, USP9X, potentially indirectly affects the stability of A/LC in cells (Tsai et al., 2017), suggesting that there may be other crucial, indirect players in this process. Therefore, we sought to utilize a more general approach rather than focusing on just the known UPP factors for A/LC turnover and screened a focused UPP inhibitor library including VCIP135 and USP9X inhibitors to identify compounds modulating BoNT/A toxicity in cells. Our initial screen utilizing mouse embryonic stem cell-derived motor neurons identified ten potential lead compounds affecting BoNT/A holotoxin mediated SNAP-25 cleavage in neurons in a pre-intoxication experimental model. Then, we examined the dose-dependent effects of the selected compounds and tested their effects on cell viability. We also explored the effects of the lead compounds on the stability and the ubiquitination of the A/LC in cells. Among the identified small molecules, PR-619 and WP1130 (degrasyn) appear to be highly crucial as they are the known inhibitors of VCIP135 and USP9X, respectively, which have been identified as DUBs affecting the half-life of A/LC (Tsai et al., 2017). Importantly, PR-619, WP1130, and Celastrol also exhibited efficacy against BoNT/A holotoxin in post-intoxication experimental conditions.

This work identified small molecule UPP inhibitors as novel modulators of BoNT/A in cells. There is an urgent need for therapeutic modalities to inhibit BoNT/A intoxication, especially after the toxin enters the neurons, where it can exhibit its toxic activity for up to 6 months (Lin et al., 2019). This is important because BoNT/A formulations are extensively used in clinics (Choudhury et al., 2021), and the wide distribution leads to concerns regarding accidents. Also, BoNT/A intoxications can naturally occur through food or liquid contaminations (Thirunavukkarasu et al., 2018). In addition, there is a potential for misuse of this relatively easy-to-produce toxin (Arnon et al., 2001). Various classes of small molecules inhibiting the toxin directly or through neuronal processes that may be curial for intoxication and/or recovery have been identified (Videnovic et al., 2014; Azarnia Tehran et al., 2015; Seki et al., 2015; Bremer et al., 2017a; Bremer et al., 2017b; Konstantinovic et al., 2018; Lin et al., 2019; Vazquez-Cintron et al., 2020; Patel et al., 2021). However, none of those candidates have yet progressed to FDA-approved drugs (Lin et al., 2019), suggesting the need for novel approaches.

One potential approach to modulate BoNT/A intoxication in cells would be targeting the persistence of A/LC (Kiris et al., 2014b). Previous work employing designer E3 ligases or A/LC specific single-chain antibodies fused to an F-box domain demonstrated that it is possible to expedite A/LC degradation in cells (Tsai et al., 2010; Kuo et al., 2011). Foreign proteins in normal functioning cells can be labeled with ubiquitin molecules for degradation by the UPP (Landré et al., 2014). However, BoNT/A can remarkably escape from this system (Tsai et al., 2017). Serotype A is the longest-lasting serotype in human botulism; it can survive within the neurons for up to 6 months (Shoemaker and Oyler, 2013). However, serotype E is active for just a few days to weeks, although it targets the same protein, SNAP-25, as BoNT/A (Shoemaker and Oyler, 2013). A recent study shows that the E3 ubiquitin ligase TRAF2 mediated ubiquitination destroys BoNT/E in neurons (Tsai et al., 2010). BoNT/A is also efficiently ubiquitinated by the E3 ligase HECTD2; however, the deubiquitinase VCIP135 removes the ubiquitin and thereby stabilizes the toxin by preventing its degradation (Tsai et al., 2017). Therefore, HECTD2 and VCIP135 activities appear to be essential for determining the lifetime of BoNT/A and the duration of its action in neurons. It is also important to mention that VCIP135 may not be the only DUB affecting A/LC half-life as it is shown that there is another DUB, USP9X, which might have an indirect effect on BoNT/A LC degradation (Tsai et al., 2017). The half-life of BoNT/A in cells can be potentially controlled if the E3 ligase and/or the DUBs can be manipulated.

Ubiquitination is a crucial post-translational process controlling many functions in cells, including protein turnover, localization, and endocytosis (Oh et al., 2018). Importantly, this is a reversible process, and DUBs can remove ubiquitin chains from the substrates (Clague et al., 2019). Many components in the UPP pathway have been targeted for drug development for various diseases, including cancer and neurodegenerative conditions. Such efforts led to the development of FDA-approved UPP targeting therapeutics. For example, proteasome inhibitors Bortezomib, Carfilzomib, Ixazomib, and E3 modulators Thalidomide, Lenalidomide, and Pomalidomide are FDA approved while there are many UPP targeting molecules, including DUB modulators, that are either in clinical trials and or in preclinical stages (Wu et al., 2020). Although proteasome and E1 targeting small molecules may lead to cellular toxicity due to non-specific effects, recent work has focused on developing specific E3 ligase and DUB modulators, and it has been shown that selective DUB inhibitor development is possible (Harrigan et al., 2018). DUBs are attractive drug targets due to their diversity and well-defined catalytic clefts (Clague et al., 2019; Basar et al., 2021), and indeed selective deubiquitinase inhibitors have been identified (Clancy et al., 2021; Varca et al., 2021).

Among the identified BoNT inhibitors in this study, PR-619 and WP1130 are of high interest as these are both DUB inhibitors with targets relevant to BoNT/A LC modulation. PR-619 is known to inhibit VCIP135 (Altun et al., 2011), which is previously identified as the main DUB inhibiting the degradation of BoNT/A LC (Tsai et al., 2017). PR-619 is a cell-permeable and reversible DUB inhibitor (Altun et al., 2011; Tian et al., 2011). It has been shown that PR-619 can enhance the ubiquitination of specific proteins, such as Bcl-2, and thereby enhance its degradation (Kuo et al., 2019), which is consistent with our findings (Figures 5, 6). WP1130, also known as degrasyn, is also a cell-permeable, reversible, small molecule that is considered the best described USP9X inhibitor (Harrigan et al., 2018). It has been shown that WP1130 can lead to enhanced ubiquitination and proteasome-dependent degradation of USP9X target proteins (Wang et al., 2014; Yang et al., 2016), which is consistent with our findings. It is, however, important to note that both PR-619 and WP1130 are not specific to the indicated targets. For example, WP1130 is considered a partially selective USP9X inhibitor, as it also targets other DUBs, USP14, USP5, and UCH37 (Kapuria et al., 2010). Similarly, PR-619 is characterized as a broad-spectrum DUB inhibitor (Altun et al., 2011). Therefore, further work is needed to determine whether other molecular players modulated by these compounds might play a role in the observed inhibition of the A/LC-mediated SNAP-25 cleavage. Importantly, there have been efforts to develop more potent and selective inhibitors than PR-619 and WP1130 targeting the same DUBs. For example, recently, a highly selective USP9X inhibitor, FT709, as compared to WP1130, was developed and characterized (Clancy et al., 2021). Future work evaluating more selective DUB inhibitors may pave the road for drug development against BoNT/A intoxication.

Another important finding of this study was identifying Celastrol as a regulator of A/LC ubiquitination and degradation (Figures 5, 6), which demonstrated efficacy against BoNT/A holotoxin mediated SNAP-25 cleavage in both pre-and post-intoxication conditions (Figures 1, 3, 7). Celastrol is a natural product derived from Tripterygium wilfordii, with anti-inflammatory and antioxidant activities (Venkatesha et al., 2016). Importantly, celastrol is pharmacologically active, and it can cross the blood-brain barrier, making it the focus of many studies in the context of many human diseases, including central nervous system disorders (Bai et al., 2021). Celastrol has multiple targets in cells (Chen et al., 2018), and its treatment has been shown to lead to degradation of specific proteins via the ubiquitin-proteasome pathway, including FANCD2 and mTOR (Metselaar et al., 2019; Wang et al., 2015; Li et al., 2018), which is in line with our results demonstrating the degradation of BoNT/A LC (Figure 5). Interestingly, Celastrol led to significant degradation of BoNT/A LC; however, although not statistically significant, we have observed an accumulation of ubiquitinated A/LC (Figure 6). It has been demonstrated that Celastrol is also a proteasome inhibitor, and therefore, it may be possible that the degree of A/LC degradation is determined by the balance between celastrol-mediated proteasome inhibition and celastrol-mediated increase in A/LC ubiquitination. A similar mechanism has been previously proposed for celastrol-mediated mTOR ubiquitination and degradation (Li et al., 2018).

Our results showed that b-AP15, NSC632839, MDBN, PYR-41, and SL-01, which do not target USP9X or VCIP135, exhibit significant effects on the degradation of BoNT/A LC at varying degrees (Figure 5). These results may be crucial because other possible mechanisms might be important for the stability of BoNT/A LC, different than VCIP135 and USP9X. Alternatively, these compounds might indirectly affect USP9X or VCIP135, leading to A/LC degradation. Notably, there has been significant research on these compounds, which may help future studies to better understand the roles of the compounds in A/LC degradation. b-AP15 targets DUBs USP14 and UCHL5 in 19S proteasome (Tian et al., 2014; Fukui et al., 2019), resulting in polyubiquitination accumulation (D'Arcy et al., 2011). However, this appears to be context-dependent as it has been shown that b-AP15 treatment can enhance poly-ubiquitination of Smad2 and Smad3 and enhance their degradation in the lysosome (Nan et al., 2016), which is consistent with our findings demonstrating that b-AP15 leads to A/LC degradation and enhances its ubiquitination in a time-dependent manner. NSC-632839 targets DUBs USP2 and USP7 and deSUMOylase SENP2 (Nicholson et al., 2008), but is also reported to induce caspase activation and apoptosis (Aleo et al., 2006). In our analyses, NSC-632839 led to increased degradation of A/LC (Figure 5); however, its effect on A/LC ubiquitination was not significant despite there was a trend of increased A/LC ubiquitination (Figure 6). P22077 is also a DUB inhibitor targeting mainly USP7 and USP47 (Altun et al., 2011). Interestingly, this compound did not exhibit a significant effect on A/LC degradation (Figure 5) and ubiquitination (Figure 6) even though it antagonized BoNT/A holotoxin mediated SNAP-25 cleavage in pre-intoxication conditions, suggesting that the compound might affect other mechanisms, such as entry steps, etc. in the intoxication mechanisms. MDBN is a cell-permeable, irreversible inhibitor of p97/valosin-containing protein that has been shown to promote ubiquitination in cells (Schweitzer et al., 2016), which is in line with our findings. PYR-41 is a cell-permeable E1 ligase inhibitor (Yang et al., 2007); however, it also targets several DUBs and some kinases (Kapuria et al., 2011). NSC-66811 and SL-01 both target E3 ubiquitin-protein ligase Mdm2 (Lu et al., 2006; Li et al., 2011). Taken together, these compounds and their potential targets within the context of BoNT/A LC stability should be further investigated in future studies. Such further characterizations should include analyses to determine whether the lead compounds might directly inhibit the enzymatic activity of A/LC, which is a zinc-dependent endopeptidase (Lebeda et al., 2010). Our analyses demonstrated that the lead compounds did not exhibit a significant effect on cell viability at tested concentrations (1.25, 2.5, 5, 10, and 20 µM), based on the MTT assay (Figure 4). Considering the experimental conditions and the exposure time of compounds in culture, in which they exhibited apparent activity against the toxin, in the initial screen (Figure 1), dose-response (Figure 3), and post-intoxication (Figure 7) analyses, we exposed the cells to compounds for 4 h in our MTT assay. However, future characterizations of the lead compounds should also include further toxicological testings with various compound treatment time points.

An effective strategy to treat botulism patients may involve a combination therapy in which specific immunoglobulins can be used to neutralize the toxin before cellular entry, coupled with drugs that can inhibit the biological activity of the already internalized toxin. Indeed, there are approved anti-BoNT antibody therapies, which are effective before BoNTs gain access into the cells (Rasetti-Escargueil and Popoff, 2019). However, as indicated above, there are currently no approved drugs against the already internalized active toxin (Lin et al., 2019). Only a limited number of BoNT molecules in cells can be sufficient to block neurotransmission (Hanig and Lamanna, 1979), and multiple groups have demonstrated that cleavage of a relatively small fraction of total SNAP-25 by A/LC is sufficient to cause significant inhibition of neuroexocytosis (Keller and Neale, 2001; Meunier et al., 2003; Keller et al., 2004; Beske et al., 2015). Given that A/LC can survive in cells for months (Shoemaker and Oyler, 2013), it would be ideal to eliminate already internalized A/LC to complement approved antibody therapies. Our work in this study focused solely on BoNT/A; however, it is crucial to develop inhibitors that can be effective against multiple BoNT serotypes. It has been shown that BoNT/E is degraded relatively rapidly compared to BoNT/A due to ubiquitination by a different E3 ligase than that of BoNT/A (Tsai et al., 2010). Future work should evaluate whether the ubiquitin-proteasome pathway can also be modulated to target serotype E and potentially other serotypes.

In summary, this study identified a subset of UPP targeting small molecules that inhibits BoNT/A LC activity in cells; two of the most efficacious compounds are DUB inhibitors. This is important as DUBs are considered druggable targets with significant clinical potential for various conditions (Magin et al., 2021). Initial studies described herein provide proof-of-concept data indicating that small molecules targeting UPP can be useful for attenuating BoNT/A intoxication. Modulating BoNT half-life in cells by small molecules can be important for research purposes to understand intoxication/recovery mechanisms and the development of effective countermeasures against botulism.