Now the findings of a new study suggest the dreaded microbe also has unexpected beneficial potential — one of its toxins can silence multiple types of pain in animals.
The research reveals that this specific anthrax toxin works to alter signaling in pain-sensing neurons and, when delivered in a targeted manner into neurons of the central and peripheral nervous system, can offer relief to animals in distress.
The work, led by investigators at Harvard Medical School in collaboration with industry scientists and researchers from other institutions, is published Dec. 20 in Nature Neuroscience.
Furthermore, the team combined parts of the anthrax toxin with different types of molecular cargo and delivered it into pain-sensing neurons. The technique can be used to design novel precision-targeted pain treatments that act on pain receptors but without the widespread systemic effects of current pain-relief drugs, such as opioids.
The need to expand the current therapeutic arsenal for pain management remains acute, the researchers said. Opioids remain the most effective pain medication, but they have dangerous side effects — most notably their ability to rewire the brain’s reward system, which makes them highly addictive, and their propensity to suppress breathing, which can be fatal.
“There’s still a great clinical need for developing non-opioid pain therapies that are not addictive but that are effective in silencing pain,” said study first author Nicole Yang, HMS research fellow in immunology in the Chiu Lab. “Our experiments show that one strategy, at least experimentally, could be to specifically target pain neurons using this bacterial toxin.”
Primed to connect
Researchers in the Chiu lab have long been interested in the interplay between microbes and the nervous and immune systems. Past work led by Chiu has demonstrated that other disease-causing bacteria can also interact with neurons and alter their signaling to amplify pain. Yet only a handful of studies so far have looked at whether certain microbes could minimize or block pain. This is what Chiu and Yang set out to do.
For the current study, they started out by trying to determine how pain-sensing neurons may be different from other neurons in the human body. To do so, they first turned to gene-expression data. One of the things that caught their attention: Pain fibers had receptors for anthrax toxins, whereas other types of neurons did not. In other words, the pain fibers were structurally primed to interact with the anthrax bacterium. They wondered why.
The newly published research sheds light on that very question.
The findings demonstrate that pain silencing occurs when sensory neurons of dorsal root ganglia, nerves that relay pain signals to the spinal cord, connect with two specific proteins made by the anthrax bacterium itself. Experiments revealed that this occurs when one of the bacterial proteins, protective antigen (PA), binds to the nerve cell receptors it forms a pore that serves as a gateway for two others bacterial proteins, edema factor (EF) and lethal factor (LF), to be ferried into the nerve cell. The research further demonstrated PA and EF together, collectively known as edema toxin, alter the signaling inside nerve cells — in effect silencing pain.
Using the quirks of microbial evolution for new therapies
In a series of experiments, the researchers found that the anthrax toxin altered signaling in human nerve cells in dishes, and it also did so in living animals.
Injecting the toxin into the lower spines of mice produced potent pain-blocking effects, preventing the animals from sensing high-temperature and mechanical stimulations. Importantly, the animals’ other vital signs such as heart rate, body temperature, and motor coordination were not affected — an observation that underscored that this technique was highly selective and precise in targeting pain fibers and blocking pain without widespread systemic effects.
Furthermore, injecting mice with the anthrax toxin alleviated symptoms of two other types of pain: pain caused by inflammation and pain caused by nerve cell damage, often seen in the aftermath of traumatic injury and certain viral infections such as herpes zoster, or shingles, or as a complication of diabetes and cancer treatment.
Additionally, the researchers observed that as the pain diminished, the treated nerve cells remained physiologically intact — a finding that indicates the pain-blocking effects were not due to injury of the nerve cells but rather stemmed from the altered signaling inside them.
In a final step, the team designed a carrier vehicle from anthrax proteins and used it to deliver other pain-blocking substances into nerve cells. One of these substances was botulinum toxin, yet another potentially lethal bacterium known for its ability to alter nerve signaling. That approach, too, blocked pain in mice. The experiments demonstrate this could be a novel delivery system for targeting pain.
“We took parts of the anthrax toxin and fused them to the protein cargo that we wanted it to deliver,” Yang said. “In the future, one could think of different kinds of proteins to deliver targeted treatments.”
The scientists caution that as the work progresses, the safety of the toxin treatment must be monitored carefully, especially given that the anthrax protein has been implicated in disrupting the integrity of the blood-brain barrier during infection.
The new findings raise another interesting question: Evolutionarily speaking, why would a microbe silence pain?
Chiu thinks that one explanation — a highly speculative one, he added — may be that microbes have developed ways to interact with their host in order to facilitate their own spread and survival. In the case of anthrax, that adaptive mechanism may be through altered signaling that blocks the host’s ability to sense pain and therefore the microbe’s presence. This hypothesis could help explain why the black skin lesions that the anthrax bacterium sometimes forms are notably painless, Chiu added.
The new findings also point to novel avenues for drug development beyond the traditional small-molecule therapies that are currently being designed across labs.
“Bringing a bacterial therapeutic to treat pain raises the question ‘Can we mine the natural world and the microbial world for analgesics?'” Chiu said. “Doing so can increase the range and diversity of the types of substances we look to in search for solutions.”
Coinvestigators included Jörg Isensee, Dylan Neel, Andreza Quadros, Han-Xiong Bear Zhang, Justas Lauzadis, Sai Man Liu, Stephanie Shiers, Andreea Belu, Shilpa Palan, Sandra Marlin, Jacquie Maignel, Angela Kennedy- Curran, Victoria Tong, Mahtab Moayeri, Pascal Röderer, Anja Nitzsche, Mike Lu, Bradley Pentelute, Oliver Brüstle, Vineeta Tripathi, Keith Foster, Theodore Price, John Collier, Stephen Leppla, Michelino Puopolo, Bruce Bean, Thiago Cunha, and Tim Hucho.
This study was funded by the Burroughs Wellcome Fund; Chan-Zuckerberg Initiative; Ipsen Pharmaceuticals; National Institutes of Health (DP2AT009499, R01AI130019, R01NS036855, NIA 5T32AG000222 fellowship, NIH NIGMS T32GM007753 fellowship), and NIH NINDS (NS111929); National Institute of Allergy and Infectious Diseases Intramural Program; European Regional Development Fund (NeuRoWeg, EFRE?0800407 and EFRE?0800408); Innovative Medicines Initiative 2 Joint Undertaking (116072-NGN-PET); and São Paulo Research Foundation (2013/08216-2 Center for Research in Inflammatory Diseases); Deutsche Forschungsgemeinschaft (271522021 and 413120531), EFRE-0800384, and LeitmarktAgentur.NRW (LS-1-1-020d).
Relevant disclosures:
S.M.L., S.P., S.M., J.M., V.T., and K.A.F. are employees of Ipsen. Chiu has received sponsored research support from Ipsen, GSK, and Allergan and is a member of scientific advisory boards for GSK and Kintai Therapeutics. This work is related to patent applications PCT/US16/49099 and PCT/US16/49106, “Compositions and methods for treatment of pain,” of which R.J.C., I.M.C., B.L.P., K.A.F., S.P., and S.M.L. are co- inventors. O.B. is a co-founder and shareholder of LIFE & BRAIN GmbH.
The FDA is increasingly approving biological drugs. In 2018, these protein-based drugs made up 25% of FDA approvals and included antibodies, growth factors, hormones, and enzymes that target a broad range of diseases [1]. The market for such drugs is expected to increase over the next few years due to their interesting properties [2].
Compared to traditional small-molecule drugs, protein- or peptide-based drugs generally show high specificity, high efficacy and high selectivity, and allow the development of drugs for a broad range of targets, notably in cancer treatment [3].
The increasing market share for biologics is even more impressive considering the inability of most of these drugs to cross the cellular plasma membrane and reach the cytosol, making their delivery a huge challenge that currently hinders the field [4]. Indeed, tools that could improve the delivery of biologics, especially those that could be applied broadly, would see immediate application and greatly benefit the drug-delivery field.
Furthermore, if delivery were sufficiently efficient, the range of potential drug targets would be drastically broadened due to the increased accessibility of intracellular proteins.
Fortunately, a naturally evolved system for delivering functional proteins into cells is provided by bacterial toxins, which have evolved to hijack cellular internalization mechanisms and developed membrane translocation devices for exactly this purpose. AB toxins are a family of bacterial toxins that include diphtheria toxin, cholera toxin, anthrax toxin, Shiga toxin, and botulinum toxin, among others [5–9].
They are named for their two components: an active part (A) that is responsible for the catalytic activity of the toxin, and a binding part (B) that is involved in binding to the cell membrane and escorting the A subunit to its destination. AB toxins have finely tuned their cellular entry to evade host defenses, providing hints and tools for protein-based drug development. In this review, we focus on how the internalization mechanisms of three AB toxins—botulinum toxin type A, anthrax toxin, and cholera toxin—can be used in different therapeutic approaches (Table 1).
We decided to focus on these three toxins based on the strong modular potential of anthrax toxin, on the already approved use of botulinum toxin type A in several neurological disorders and on the wide variety of therapeutical applications of cholera toxin. The practical approaches presented in this review take advantage of both the intrinsic properties of the toxins as well as the modularity of both the A and B subunits, all aspects that can be further extended to other AB toxins.
Anthrax Toxin
Anthrax Toxin Internalization Mechanism
An important concern for animal health and human public safety in the context of bioterrorism, anthrax toxin is an AB toxin produced by the gram-positive spore-forming bacterium Bacillus anthracis. This toxin consists of a B subunit, protective antigen (PA), and two catalytic A subunits, lethal factor (LF) and edema factor (EF). PA is an 83-kDa protein
that is responsible for the binding of the toxin to its main receptors, capillary morphogen- esis 2 (CMG2) and tumor endothelial marker 8 (TEM8) [12,13]. LF is an 91-kDa matrix metalloprotease that cleaves the MAPKK family members, which impairs the associated signaling pathways and eventually leads to apoptosis, especially in macrophages [29,80].
EF is a calmodulin-dependent adenylyl cyclase that increases the cytosolic cAMP lev- els. This review briefly describes the internalization process of anthrax toxin and, for a more in-depth understanding of this mechanism, readers are oriented towards previously published reviews [6].
Initially in LF and EF internalization, extracellular PA binds to one of its receptors, CMG2 or TEM8, and then is cleaved by furin-family proteins (Figure 1B). This cleavage allows PA to oligomerize into heptamers or octamers, also called pre-pores [15,16,81], which can then recruit three or four LF or EF subunits, respectively, for internalization. On the cytosolic side, PA binding to the TEM8 or CMG2 receptor causes it to release from the actin cytoskeleton [82,83], allowing ubiquitination of the receptor, which triggers endocytosis of the receptor-anthrax toxins complex [82].
Anthrax toxin and its receptors are then targeted to early endosomes where they are sorted in endosomal intraluminal vesicles (ILVs) and trafficked through the endocytic pathway towards late endosomes [21]. On the way to late endosomes, the acidification of the microenvironment induces a conformational change in the PA pore [25], and this low pH is also required for the translocation of LF [26]. Pores can form at the limiting membrane of the endosomes, translocating LF or EF directly into the cytosol, though most pores form in the membrane of ILVs [21,84].
These pores allow the translocation of LF or EF to the lumen of ILVs and, by back-fusion of ILVs with the limiting membrane of late endosomes, LF or EF eventually reaches the cytosol [21]. In opposition to BoNT/A, evidence suggest that LF has a very short half-life in the cytosol and its long-term effect relies on its ability to remain dormant in ILVs which stochastically back-fuse with the membrane of endosomes over a long period of time [21].
Anthrax Toxin Therapeutic Applications
The therapeutic potential of anthrax lethal toxin was originally exploited in anti-cancer treatments due to its inhibitory effect on the MAPKK-associated pathway. Unlike normal cells, cancer cells usually rely on only a few dysregulated pathways to increase their growth, survival, or motility.
Accordingly, some cancers, such as melanoma bearing the V600E BRAF mutation, mostly rely on the constitutively activated MAPK pathway for cell growth and survival, and anthrax toxin was shown to decrease both these processes in this particular cell line [32]. Similarly, anthrax lethal toxin was shown to reduce cell growth and tumor angiogenesis in renal cell carcinoma and to reduce cell motility and invasiveness in astrocytes by targeting the MAPK pathway [33].
Although anthrax lethal toxin showed interesting intrinsic anti-tumor properties, most of its potential in therapy relies on its modular properties, like its ability to translocate different non-native proteins, drugs, and other molecules. As mentioned previously, PA oligomers create a pore in endosomes, allowing LF to eventually reach the cytosol, suggesting that LF fusion proteins could go through the pore as well—as long as they can successfully unfold while passing through the pore and refold later in the cytosol.
In the 1990s, the first attempts to fuse proteins to the N-terminus of the LF subunit were done to target proteins to the cytosol and confirm the potential of anthrax toxin as a delivery system. FP59, a fusion between the N-terminus of LF (LFN) with the ADP-ribosylation domains of Pseudomonas exotoxin A, was the first successful translocation of a foreign protein into the cytosol [39].
Shortly after, both catalytic domains of the Shiga and diphtheria toxins reached the cytosol when fused to LFN, further supporting that the N-terminal residues of LF were sufficient to translocate complicated polypeptide chains through the PA pore [40,41]. However, Blanke et al. later showed that a simple positively-charged polycationic peptide could replace LFN for the delivery of diphtheria toxin to the cytosol [42].
Besides bacterial toxins, the LFN delivery system was shown to be useful in other applications, such as the development of a potential HIV vaccine and the treatment of neurodegenerative diseases [43,44]. In a broader perspective, Rabideau et al. assessed the feasibility of translocation through the PA pore for many different cargo molecules, from short or cyclic peptides to small molecule drugs.
They concluded that while non- canonical peptides and small-molecule drugs, such as doxorubicin, can be translocated, cyclic peptides and the small molecule docetaxel cannot, which they hypothesized was due to rigidity of the cargo [45]. These examples provide strong evidence that many different cargo proteins can be delivered to the cytosol both in vitro and in vivo using anthrax toxin, which can be used for the targeted delivery of vaccines, drugs, and other proteins.
Besides its ability to translocate different non-native cargos, another modular char- acteristic of PA lies in the specificity of the protease that processes it, thereby allowing it to oligomerize. In the last two decades, several groups focused on unraveling the best combinations of mutations in PA that would allow more targeted and less toxic tumor ther- apies.
The two PA mutants, PA-L1 and PA-U2, were programmed to be specific for several cancer cell lines in vitro by changing the cleavage site from furin to matrix metallopro- teases (MMPs) and urokinase plasminogen activator (uPA), respectively [52,53], which are overexpressed in many cancer types while not very abundant at the surface of normal cells. In particular, PA-U2 showed a strong anti-tumor activity and specificity when combined with FP59 in mice [54].
To make the tumor targeting more specific, PA-L1 and PA-U2 were mutated on their homo-oligomerization domain to render them complementary, making them even more specific to cancer cells expressing both proteases. This approach was shown to be efficient with different sets of PA mutants both in vitro and in vivo [85–88].
In addition to their use as anti-tumor drugs, the protease-specific PA mutants were also used in combination with radioactively labelled LF or LFN-β-lactamase fusion protein to develop methods of imaging plasma membrane protease activity in tumors or in cancer cell lines, respectively [89,90].
Using the potential of PA to internalize molecules, several research groups adapted this technology to allow cancer-specific receptors to bind and internalize PA-fusions specific for those receptors. Varughese et al. were the first to unravel the potential of this strategy by targeting FP59 to a c-Myc-specific 9E10 hybridoma cell line using a PA-c-Myc fusion protein [55]. McCluskey et al. used a similar approach containing a mutated PA (mPA) that cannot bind its natural receptors fused with a high-affinity Affibody, ZHER2, targeting the HER2 receptor [56].
They showed that both mPA-EGF and mPA-ZHER2 could deliver an LFN-fused diphtheria toxin catalytic domain (DTA) to kill several cancer cell lines depending on the presence of their respective receptors [56]. Based on these observations, PA can form pores and deliver cargos as long as the targeted receptor is able to internalize, broadening the number of potential targets at the cell surface of cancer cells.
Additionally, Loftis et al. used an mPA fused with the single-chain variable fragment (scFv) of an antibody to internalize and deliver LFN-DTA through EGFR or carcinoem- bryonic antigen, which could kill pancreatic cancer cells overexpressing the two receptors at the plasma membrane. For additional specificity towards their pancreatic cancer cell line, they made an LF-RRSP fusion protein which targets the Ras–ERK signaling pathway, crucial for many pancreatic cancer cells [57].
Similarly, Becker et al. used designed ankyrin repeat proteins (DARPins) fused to a PA-CMG2-based construct to specifically target trans- membrane glycoprotein epithelial cell adhesion molecule (EpCAM) at the surface of cells. These engineered constructs were shown to target EpCAM-expressing cells with a high specificity and to deliver LFN-based constructs to the cytosol [58].
Overall, these engineered proteins show that both the A and B subunits of anthrax toxin have strong potential as a protein delivery system, and they open many new routes for investigating the development of therapeutics. However, the immunogenicity of anthrax toxin subunits, as illustrated by the use of PA in anthrax vaccines, for example, remain a challenge to address in its therapeutical applications [91].
