METHODS OF USE THEREOF

PRIORITY

This application claims the benefit of U.S. Provisional Application No. 62/376,960 filed August 19, 2016, which is hereby incorporated by reference in its entirety.

FIELD OF THE INVENTION

The present disclosure relates to methods and compositions for neutralizing toxins. In particular, the present invention provides bacteriophages for neutralizing bacterial toxins, and their use for treating, preventing, and controlling bacterial toxicosis in humans and animals.

BACKGROUND OF THE INVENTION

Bacterial infections are often associated with the release of bacterial toxins that are responsible for the manifestation of the disease ("toxicoses"). While many bacteria release virulence factors that enhance the ability of the bacteria to survive in the infected host, for example, by subverting or avoiding the immune system, obtaining nutrients from the host, and inducing dormancy to survive host defenses and antibiotic treatment; for bacterial toxicoses, the disease can be recapitulated by the toxin itself in the absence of any bacteria.

Bacterial toxicoses include intestinal diseases caused by bacterial contamination of water sources across the world. These include cholera caused by Vibrio spp. and cholera toxin, as well as E. coli and Shigella diarrheal disease (e.g., shiga toxin). In addition, childhood immunizations against diphtheria and tetanus toxoids are part of routine immunizations in most of the world. Bacterial toxicoses are also relevant to the food industry. For example, necrotic enteritis in poultry is largely caused by Clostridium perfringens toxin. C. perfringens is a commensal bacteria found in about 75% of chickens.

While vaccination is a useful strategy for some toxoid-based diseases, some bacterial diseases do not have useful vaccines or traditional vaccination is not practical from either a logistic or economic perspective. In addition, vaccination must be done prophylactically prior to infection and in general is not a useful strategy for treating active infections. Active infections can sometimes be treated by either antibiotic therapy and/or passive immunization with antibodies targeted to the bacteria or bacterial toxin.

Accordingly, there remains a need for prevention, treatment, and control of bacterial toxicoses, including for improvement in human and animal health. The following disclosure addresses this and other needs.

SUMMARY OF THE INVENTION

In one aspect, the present disclosure provides bacteriophages that encode toxin- binding proteins for expression in host bacterial cells. Host cells include toxin- producing bacteria, or commensal bacteria that do not themselves produce the target toxin. The bacteriophage provides for the neutralization of toxin, including those produced by bacterial pathogens in humans and animals, such as Clostridium difficile (C. difficile), Clostridium perfringens (C. perfringens), Corynebacterium diphtheriae, Clostridium tetani, Clostridium botulinum, Vibrio cholerae, Enterotoxigenic E. coli, Shigella dysenteriae type 1 {Shiga bacillus), Staphylococcus aureus, Streptococcus pyogenes and Streptococcus pneumoniae. In certain embodiments, the toxin is produced by a bacterial pathogen, including a commensal bacterial species that becomes a pathogen due to pre-disposing conditions, for example, in the gastrointestinal tract, upper respiratory system, or nasopharyngeal tract.

The toxin-binding protein may take a variety of forms, including antigen- binding portions of antibody molecules, or engineered antibody platforms or antibody mimics known in the art. An exemplary toxin-binding molecule is a single-chain variable fragment antibody (scFv).

The toxin-binding protein in some embodiments is directed against or neutralizes toxin A or toxin B from C. difficile. In some embodiments, the toxin is a C. perfringens toxin, such as beta-like toxin B. C. perfringens is a significant pathogen of poultry. In some embodiments, the toxin-binding protein neutralizes pneumolysin, a toxin produced by S. pneumoniae during pneumococcal pneumonia.

The bacteriophage may be engineered from a lysogenic or lytic bacteriophage. For example, in some embodiments the bacteriophage is a lytic phage that targets/infects the toxin-producing bacteria. In some embodiments, the bacteriophage is a lytic phage that targets/infects a commensal bacteria, which may or may not produce the toxin. In some embodiments, the bacteriophage is a lysogenic phage that targets a commensal bacteria that does not itself produce the target toxin. Exemplary bacteriophage include those of the order Caudovirales, such as Siphoviridae, Myoviridae and Podoviridae. The toxin-binding protein may be released upon lysis of infected bacteria, or in addition or alternatively, release of the toxin-binding protein is directed by an encoded secretory signal. Alternatively, the toxin-binding protein may be fused with a phage structural protein or other protein.

In another aspect, the present disclosure provides compositions suitable for use as therapeutics for treating, preventing, or controlling bacterial toxicosis, including C. difficile associated disease in humans, or necrotic enteritis in poultry, among others. In some embodiments, the compositions are suitable for use in treating or controlling pneumococcal pneumonia. The compositions may comprise a carrier for delivering the bacteriophage to the site of infection, such as the GI tract or nasopharyngeal tract or middle ear. With regard to animal health, bacteriophage may be mixed or applied to feed or drinking water, providing an economical means for treating large flocks or herds.

In another aspect, the present disclosure is related to methods of preventing, ameliorating, treating, or controlling a toxin-associated infection, comprising administering to the subject a bacteriophage described herein. The subject may be a human or animal subject, and may have been exposed, or may be exposed, to toxin producing bacteria or spores of such bacteria, such as C. difficile or C. perfringens in some embodiments.

Other aspects and embodiments will be apparent from the following detailed description.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1 shows phage plaques resulting from transformation of genomic DNA isolated from natural GRCS phage, or engineered phage DNA assembled in a bacterial artificial chromosome vector, on Sau strain RN4220 (left panel), and amplification of DNA isolated from those phage plaques by PCR (right panel).

Figure 2 is a schematic showing sites of GFP gene insertion into the phage genome, assembled using a bacterial artificial chromosome vector. Figure 3 shows that insertion of GFP gene at locus 1 results in only natural phage.

Figure 4 shows that insertion of GFP at locus 2 shows some GFP gene insert in recovery lysate.

Figure 5 shows that 0.5M sucrose stabilizes phage carrying the GFP gene in recovery lysate at locus 2.

DETAILED DESCRIPTION

The present disclosure provides bacteriophages that encode toxin-binding proteins for expression in host bacterial cells, as well as methods for treating, preventing, or controlling bacterial toxicoses. Bacteriophages (phage) are viruses that specifically infect bacteria, and not human or animal cells. The bacteriophages according to this disclosure express and optionally secrete anti-toxin molecules, including but not limited to single-chain antibodies, which will bind, sequester, and/or neutralize target toxins.

In some embodiments, the toxin-binding protein binds to a toxin from a bacterial species, which in some embodiments is a human or animal pathogen, or a commensal bacteria of poultry, livestock or humans. In some embodiments, the pathogen is of the genus Clostridium, Vibrio, Escherichia, Streptococcus, Staphylococcus, or Shigella. In these or other embodiments, the toxin-producing bacterium is selected from Clostridium difficile (C. difficile), Clostridium perfringens (C. perfringens), Corynebacterium diphtheriae, Clostridium tetani, Clostridium botulinum, Vibrio cholerae (e.g., 01 or 0139), Enterotoxigenic E. coli, Shigella dysenteriae, Staphylococcus aureus, Streptococcus pyogenes, and Streptococcus pneumoniae. In further embodiments, the toxin-binding protein binds to a toxin from bacterium including, but not limited to, those disclosed in U.S. Patent Application No. 2009/0087478, which is hereby incorporated by reference in its entirety.

In various embodiments, the target toxin is a Clostridium toxin. Examples include a Clostridium difficile toxin, such as toxin A or toxin B; or a Clostridium perfringens toxin, such as toxin types A, B, C, D, or E, or combination thereof, or beta- like toxin B. C. perfringens toxins include C. perfringens alpha toxin (CPA), C. perfringens beta toxin (CPB or CPB2), epsilon toxin (ETX), iota toxin (ITX), perfringolysin O (PFO), C. perfringens enterotoxin (CPE), toxin perfringens large (TpeL), and Necrotic enteritis toxin B-like (NetB), as further described in Ural et al., Future Microbiol. 9(3), 361-377 (2014), which is hereby incorporated by reference in its entirety.

In some embodiments the toxin is pneumolysin, for example, as produced by

Streptococcus pneumoniae during pneumococcal pneumonia.

The toxin-binding protein may be an antibody or antigen-binding portion thereof. Various antibody variants and antigen-binding platforms are known, including those for preparing small, compact, target-binding molecules. Examples include a single-domain antibody, a heavy-chain-only antibody (VHH), a single-chain variable fragment antibody (scFv), a shark heavy-chain-only antibody (VNAR), a microprotein (cysteine knot protein, knottin), a DARPin, Adnectin, Tetranectin, Affibody; Transbody, Anticalin, Affilin, Microbody, peptide aptamers, phylomer, stradobody, maxibody, evibody, fynomer, an armadillo repeat protein, a Kunitz domain, avimer, atrimer, probody, immunobody, triomab, troybody, pepbody, UniBody, and DuoBody. Exemplary antigen-binding formats are described in US Patent Nos. or Patent Publication Nos. US 7,417, 130, US 2004/132094, US 5,831,012, US 2004/023334, US 7,250,297, US 6,818,418, US 2004/209243, US 7,838,629, US 7, 186,524, US 6,004,746, US 5,475,096, US 2004/146938, US 2004/157209, US 6,994,982, US 6,794, 144, US 2010/239633, US 7,803,907, US 2010/1 19446, and/or US 7, 166,697, the contents of which are hereby incorporated by reference in their entireties.

In some embodiments, the toxin-binding protein is a single-chain antibody. A single-chain antibody is a modified variable heavy and/or light chains of human, camelid, or avian antibodies, which can be expressed as single domain proteins recognizing specific antigens. These single-chain antibodies generally lack immune effector function, as they do not contain the effector function domains of typical IgG from mammalian systems. For example, single-chain variable fragment antibodies (scFv) binding to the desired toxin target can be identified based on phage display, where the scFv is covalently attached to the phage particle, and screened for binding against specific antigens. See Z. A. Ahmad, et al. Clin Dev Immunol, vol. 2012, p. 980250, 2012; I. Benhar, et al, JMol Biol, vol. 301, pp. 893-904, Aug 25 2000; A. E. Nixon, et al, MAbs, vol. 6, pp. 73-85, Jan-Feb 2014; A. M. Shukra, et al. Eur J Microbiol Immunol (Bp), vol. 4, pp. 91-8, Jun 2014, all of which are hereby incorporated by reference in their entirety. In addition to binding to the antigen target, scFv can also be screened in functional assays for inhibition of target toxin function.

In some embodiments, the toxin-binding protein has a molecular weight of less than about 30 kDa, or less than about 20 kDa, or less than about 15 kDa, or less than about 10 kDa, or less than about 5 kDa. In various embodiments, the toxin-binding protein binds to its target with an affinity (KD) of better (less) than about 100 mM, or less than about 10 mM, or less than about 1 mM, or less than about 100 M, or less than about 10 μΜ, or less than about 1 μΜ, or less than about 100 nM, or less than about 10 nM, or less than about 1 nM in various embodiments.

In embodiments that employ scFv antibodies, the toxin-binding protein can comprise one or several (2, 3, or 4) VH and VL hypervariable region chains (the portion of each chain that together form the antigen-binding epitope) linked together in head to head or head to tail configurations by short peptide linkers.

More specifically, in some embodiments, a library of scFv antibodies (or other binding-protein format described above) is provided using a phage, yeast, ribosome or other display technology, which is useful for screening for binding proteins to a selected bacterial toxin antigen. For example, in the case of C. dfficile, the target may be toxin B (tcdB), or alternatively toxin A (tcdA). For C. perfringens, the target may be toxin A (pic) and/or β-like toxin (netB). Once the scFv antibody is selected, either from direct screening or molecular evolution to enhance affinity to the toxin antigen, the gene encoding the scFv antibody is then inserted into a phage genome. In some embodiments, the phage may be a lytic phage of the family Podoviridae, such as phi24R, phiZP2, phiCPV4 or phiCP7R for C. perfringens, or similar phage. In some embodiments, the phage is a lytic phage of the P68 genus of the family Podoviridae such as GRCS, P68, 44AHJD, SAP-2, S24-1, 66 phage which targets Staphylococcus aureus. In some embodiments, the phage is Cp-1 or SOCP, or other member of the Picovirinae family that infects Streptococcal pneumoniae. The phages can have broad spectrum activity against target bacterial field or clinical isolates, which may be tested prior to the initiation of genome engineering. For example, in the case of C. difficile, since it lacks identified lytic phages, temperate phages could be converted to lytic phages by engineering the phage to interfere with mechanisms of genomic integration. Alternatively, phage that infect common commensal organisms may be engineered in similar fashion. In further embodiments, lytic phages to C. difficile may be obtained by screening (e.g., screening of stool samples). In some embodiments, the phage is a lytic phage that infects commensal bacteria, which may or may not produce the toxin. In these embodiments, the phage may be engineered to be weakly lytic.

In some embodiments, phage lysis is controlled by the expression and concentration-dependent oligomerization of holin to form a pore, which allows lysin to exit the cytoplasm and access the cell wall for degradation and lysis. By mutating the holin gene of the phage, the oligomerization rate can be reduced and lysis delayed. Thus, in some embodiments, the holin and/or lysin genes are mutated to delay lysis.

The bacteriophage engineered to encode a toxin-binding protein may be produced according to techniques known in the art (See, US Patent Nos. 8,182,804, 9,056,899, and 8,153, 119; and U.S. Patent Application Publication No. 2015/0050717, all of which are incorporated herein by reference in their entirety). In some embodiments, the bacteriophage is produced by inserting into the phage genome the nucleic acid sequence encoding the toxin-binding protein. The selected scFv or other toxin-binding protein may be either covalently attached (e.g., by translational fusion) to a phage protein, designed for secretion from the bacteria during phage infection, or released from phage-infected bacteria upon lysis. For example, using genome assembly from PCR fragments from isolated phage linear double stranded DNA, the phage genome may be assembled in vitro, or by homologous recombination in yeast vectors, into a workhorse plasmid-based construct, such as a bacterial artificial chromosome (BAC) or yeast artificial chromosome (YAC). Transformation of the assembled genome in the BAC or YAC may then be transformed into the target bacterial species and phage production assessed by plaque formation in a standard overlay assay. In additional embodiments, once phage production is confirmed, the nucleic acid sequence encoding toxin-binding protein may be inserted into the phage genome in several different loci to determine expression levels using a combination of PCR detection of recovered engineered phage (e.g. for stability of the inserted gene) and immunodetection of epitope -tagged versions of the scFv genes to determine protein expression. The nucleic acid sequence encoding toxin-binding protein described herein may contain a secretion sequence to promote secretion of the toxin-binding protein during phage infection; it may be expressed independently of secretion and released upon lysis of the phage-infected cell; or it may be fused to a capsid protein or other structural phage protein, and released as a phage/anti-toxin particle that sequesters the targeted toxin.

In some embodiments, the phage is engineered for stability of the transgene, which can include positive selection in the presence of serially decreasing osmotic stabilizer (such as sucrose), and/or negative selection of phage without the transgene (e.g., by targeting CRISPR/Cas9 endonuclease to the wild type locus).

In some embodiments, the engineered phage is tested for anti-toxin activity in lysed culture supernatants of engineered phage, e.g., by ELISA and/or western blot against the target toxin. In addition, the lytic activity of the engineered phage against a number of pathogen clinical or field isolates may compared to the spectrum of the original natural phage chosen for engineering. The spectrum may be enhanced by identifying the recognition baseplate proteins in the chosen phage genomes, and modification of these baseplate proteins by either baseplate swaps with other phage baseplates, or modification such as with other possible recognition elements (e.g. linear amphipathic cationic antimicrobial membrane-binding peptides).

In some embodiments, the bacteriophage is a virulent phage. The bacteriophage may be identified or engineered from a lysogenic bacteriophage. For example, a virulent phage may be created from the lysogenic bacteriophage by deletion of genes in the phage genome that are required for lysogeny. In some embodiments, the phage is of the family Podoviridae and subfamily Picovirinae. Examples include GRCS phage, which infects Staphylococcus aureus. In some embodiments, the phage (e.g., Cp- 1 or SOCP phage) infects Streptococcus pneumoniae. Other families of phage, which may be engineered in accordance with this disclosure, include Myoviridae and Siphoviridae.

In some embodiments, the bacteriophage is a Podoviridae that contains a deletion of all or part of the minor tail protein. Phage genomes are small, and tightly packed, with only around 20 to 22 open reading frames. While most of the genes are not believed to be dispensable, some variants of Podoviridae, for example, exhibit truncated forms of the minor tail protein, suggesting that this open reading frame might be partially or completely deleted to make room for the gene encoding the toxin- binding protein. In some embodiments, this will reduce or alter the host range of the phage. In some embodiments, the toxin is a Clostridium toxin, and is optionally a Clostridium difficile toxin. In these or other embodiments, the bacteriophage may also infect C. difficile. C. difficile is a gram-positive, motile, obligate anaerobic, spore- forming bacterium. It is the causative agent of CDAD {Clostridium difficile associated disease) in humans, usually predisposed by antibiotic therapy in hospital and long-term care facilities. It may cause severe colitis, sepsis and death. The bacterium is acquired through ingestion of spores, usually by the fecal-oral route, and can be transmitted person to person.

In some embodiments, the target toxin to be neutralized is toxin A or toxin B from C. difficile. Toxin A {tcdA) and toxin B {tcdB) are chromosomally encoded at the PaLoc (pathogencity locus) that encodes a number of genes besides the toxin genes, and which contribute to virulence and pathogenicity. There are 34 known variants in the PaLoc, as determined by RFLP, including SNPs largely in tcdB, in either the catalytic or receptor binding domains, and including insertions and deletions (mostly in tcdA). Despite the diversity of toxinotypes, tcdA and tcdB are generally highly homologous to each other (M. Rupnik and S. Janezic, J Clin Microbiol, vol. 54, pp. 13- 8, Jan 2016). There is a third toxin designated binary toxin (CDT), with a less prominent role in pathogenesis (D. N. Gerding, et al., Gut Microbes, vol. 5, pp. 15-27, Jan-Feb 2014). Toxin A and toxin B mechanism of action is to glucosylate the Rho- family of GTPases, inducing actin depolymerization (I. Just, et al., Nature, vol. 375, pp. 500-3, Jun 8 1995; I. Just, et al., J Biol Chem, vol. 270, pp. 13932-6, Jun 9 1995). The general structure of these large clostridial toxins follows the ABCD toxin model, where the enzymatic activity of the toxin (A) resides in the N-terminal domain, the receptor binding domain (B) resides in the C-terminal domain, the auto-proteolytic cutting (C) and the short hydrophobic pore-forming domain for cytosolic delivery (D) are between the A and B domains (T. Jank and K. Aktories, Trends Microbiol, vol. 16, pp. 222-9, May 2008).

The use of antibodies to neutralize toxin effects in CDAD is exemplified by actoxumab (anti-toxin A) and bezlotoxumab (anti-toxin B) as shown in U.S. Patent Nos. 8,236,311, 7,625,559, 8,257,709, 8,609,111, 9,217,029, all of which are incorporated herein by reference in their entirety. Bezelotoxumab (Zinplava™) showed significant reduction of recurrence of CDAD in combination with standard of care oral antibiotics. Single-chain antibodies have also been developed to clostridial toxins for potential use in diagnostics (X. K. Deng, L. A. Nesbit, and K. J. Morrow, Jr., Clin Diagn Lab Immunol, vol. 10, pp. 587-95, Jul 2003) or therapeutics (G. Hussack, et al., J Biol Chem, vol. 286, pp. 8961-76, Mar 18 2011). Accordingly, toxin-binding proteins can be engineered to contain antigen-binding sequences of actoxumab or bezlotoxumab. Optionally, the toxin-binding protein comprises an antigen-binding sequence of actoxumab or bezlotoxumab, optionally from 1 to 10 (e.g., 1 to 5) amino acid substitutions, insertions, or deletions. In some embodiments, the toxin-binding protein is designed to bind to a competing epitope to actoxumab or bezlotoxumab.

The bacteriophage infecting C. difficile may be identified or engineered from a lysogenic bacteriophage. In some embodiments, a virulent phage against C. difficile may be created from the lysogenic bacteriophage by deletion of genes in the phage genome that are required for lysogeny. For example, the lysogenic phages that infect Clostridium difficile have been described (J. Y. Nale, et al, Antimicrob Agents Chemother, vol. 60, pp. 968-81, Feb 2016; K. R. Hargreaves and M. R. Clokie, Front Microbiol, vol. 5, p. 184, 2014; US Patent Application Publication No. 2015/0290263, which are all hereby incorporated by reference in their entirety). The lysogenic bacteriophage include NCTC 12081404, NCTC 12081405, NCTC 12081406, NCTC 12081407, NCTC 12081408, NCTC 12081409, and NCTC 12081410. Generally, the bacteriophage may be engineered from a phage selected from Caudovirales, Siphoviridae, Myoviridae and Podoviridae, for example.

In additional embodiments, the bacteriophage is against a commensal bacterium other than C. difficile. For example, the nucleotide sequence encoding the toxin-binding protein may be inserted into a lytic phage that infects a commensal bacterium. The commensal bacterium may be of a species of Clostridiales, Lachnospiraceae, Ruminococcaceae, Veillonellaceae, Bacteroidales, Lactobaci Hales, Clostridium, Bacteroides, Fusobacterium, Streptococcus, Staphylococcus, Enterococcus, Parabacteroides, or Salmonella. For example, the commensal bacterium may be Escherichia coli or Lactobacillus spp. , commonly found in CDAD stool samples (J. Y. Chang, et al, J Infect Dis, vol. 197, pp. 435-8, Feb 1 2008; M. J. Hopkins and G. T. Macfarlane, J Med Microbiol, vol. 51, pp. 448-54, May 2002; M. J. Hopkins, R. Sharp, and G. T. Macfarlane, Gut, vol. 48, pp. 198-205, Feb 2001; A. Khoruts, et al., J Clin Gastroenterol, vol. 44, pp. 354-60, May-Jun 2010). In some embodiments, the commensal bacterium is Streptococcus (e.g., Streptococcus pneumoniae), which is commonly found in the upper respiratory system (e.g., nasopharyngeal tract), but can become pathogenic. In some embodiments, the bacteriophage is a phage that is deficient in lysis, but does not integrate its genome into the genome of the host bacteria. The phage may accumulate in the cytoplasm of the infected commensal bacteria, while expressing the anti-toxin and transporting it through a secretion pathway. The secreted anti-toxin may then be available to sequester toxin produced by C. difficile. In some embodiments, the antitoxin molecule is not secreted, but sequesters toxin inside the pathogenic bacteria (e.g., S. pneumoniae).

In some embodiments, the target toxin is produced by Clostridium perfringens (C. perfringens). In these embodiments, the bacteriophage may infect C. perfringens, or may target another commensal bacteria. The bacteriophage may be engineered from a phage selected from Caudovirales, Siphoviridae, Myoviridae and Podoviridae, for example. C. perfringens is a commensal gram-positive facultative anaerobic spore- forming bacterium, found in the intestinal tract of poultry (e.g., in about 75 to 80% of chickens). It is the causative agent of necrotic enteritis in chickens, primarily through the mechanism of action of secreted toxins (J. F. Prescott, et al., Avian Pathol, vol. 45, pp. 288-94, Jun 2016). C. perfringens consists of 5 toxinotypes (A-E), each toxinotype secreting a different combination of toxins. Type A is the most common cause of necrotic enteritis in poultry, producing primarily toxin A (pic), with type C less common, but producing both toxin A and toxin B (F. Van Immerseel, et al., Avian Pathol, vol. 33, pp. 537-49, Dec 2004). However, most notably the beta-like toxin B (netB) has been shown to be a primary toxin causing the necrotic enteritis in poultry (A. L. Keyburn, et al, PLoS Pathog, vol. 4, p. e26, Feb 8 2008; A. L. Keyburn, et al., Toxins (Basel), vol. 2, pp. 1913-27, Jul 2010; K. A. Hassan, et al., Res Microbiol, vol. 166, pp. 255-63, May 2015; A. L. Keyburn, et al., Infect Immun, vol. 74, pp. 6496-500, Nov 2006; F. A. Uzal, et al., Future Microbiol, vol. 9, pp. 361-77, 2014). Toxin A may be primarily involved in maintaining sub-clinical infection and in myonecrosis in poultry. Predisposing conditions that initiate overgrowth of the organism and induce necrotic enteritis include: stress on the animals and intestinal mucosal damage caused by parasitic infection (e.g., apicomplexan Eimeria spp. parasites). Intestinal damage caused by parasitic infection may release nutrients into the gut, prompting overgrowth of C. perfringens. Similarly, diets consisting of high indigestible, water-soluble non- starch polysaccharides, such as rye, wheat and barley, predispose chickens to necrotic enteritis. This may be due, in part, to the presence of fungal mycotoxins, such as fumonisin B, that have been shown to induce intestinal epithelial damage and are common contaminants in poultry feed (G. Antonissen, et al, Vet Res, vol. 46, p. 98, 2015).

In some embodiments, the toxin to be neutralized is toxin beta-like toxin B

(netB) from C. perfringens. NetB is a beta-like pore-forming toxin similar to the Staphylococcus aureus a-hemolysin, γ-toxin and leukocidin toxin. These toxins self- assemble and target host epithelial membranes, forming beta-barrel pore complexes in the membrane leading to cytosolic content leakage and loss of membrane polarity (A. L. Keyburn, et al., Toxins (Basel), vol. 2, pp. 1913-27, Jul 2010). NetB is plasmid encoded and therefore represents a mobile genetic element for toxin transfer by conjugation. The NetB pore complex consists of six monomers forming a β-barrel pore divided into three ultrastructural domains, the β-sandwich, rim, and stem (C. G. Sawa, et al., J Biol Chem, vol. 288, pp. 3512-22, Feb 1 2013). The rim domain mediates binding of the complex to phospholipid membranes. Moreover, cholesterol has been shown to play a role in monomer oligimerization to form the pore structure, and thus monomer produced by the bacteria should be available for sequestration and inhibition of oligomer formation.

The bacteriophages may be lytic bacteriophages against C. perfringens (US Patent Nos. 9,320,795, 7,625,740, and 7,625,739, U.S. Patent Application Publication No. 2016/0076003, and 2016/0076004, C. A. Morales, et al., Arch Virol, vol. 157, pp. 769-72, Apr 2012, B. B. Oakley, et al, BMC Genomics, vol. 12, p. 282, 2011, B. S. Seal, Poult Sci, vol. 92, pp. 526-33, Feb 2013, B. S. Seal, et al , Arch Virol, vol. 156, pp. 25-35, Jan 2011, N. V. Volozhantsev, et al. , PLoS One, vol. 7, p. e38283, 2012, N. V. Volozhantsev, et al, Virus Res, vol. 155, pp. 433-9, Feb 2011, all of which are hereby incorporated by reference herein in their entirety). The phages cover three families of Caudovirales, the Siphoviridae, Myoviridae and Podoviridae. These families are characterized by related ultrastructure of the virion particles, and not necessarily by genomic sequences. Phylogeny based upon genome relatedness is progressing as more phages are characterized and sub-families are being characterized through genome sequencing. Podoviridae sequences for virulent phages to Clostridium perfringens have been deposited in the NCBI database (C. A. Morales, et al, Arch Virol, vol. 157, pp. 769-72, Apr 2012; N. V. Volozhantsev, et al., PLoS One, vol. 7, p. e38283, 2012). The Podoviridae phages infecting C. perfringens are lytic phages with small genomes and are similar in structure and organization to the S. aureus phage GRCS. The toxin-binding protein gene, selected by phage, yeast or other display technology, would be engineered into the C. perfringens phage genome according to these embodiments.

In some embodiments, the encoded toxin-binding protein further comprises an

N-terminal secretory signal, directing secretion of the toxin-binding protein from the host bacteria into the gut environment. Exemplary secretory signal sequences are disclosed, for example, in U.S. Patent Publication No. 2015/0050717, which is hereby incorporated by reference in its entirety. Alternative or in addition, the toxin-binding protein is released into the gut environment upon lysis of the engineered phage-infected bacterium. Alternatively, the toxin-binding protein is fused to a structural phage protein (e.g. capsid protein) or a lytic enzyme.

In another aspect, the bacteriophage comprises a promoter sequence operatively linked to direct expression of the anti-toxin protein gene disclosed herein. In some embodiments, the promoter is a bacteriophage promoter. In some embodiments, the toxin-binding protein is expressed as part of an operon, or individually. Additional promoters that may be used are disclosed, for example, in U.S. Patent Publication No. 2010/0322903 (which is hereby incorporated by reference in its entirety), and at partsregistry.org/cgi/partsdb/pgroup.cgi?pgroup=other_regula tor&show=l.

The bacteriophage may be included as part of a composition for delivering the phage to its intended site of action, such as the intestinal tract. Compositions in some embodiments may contain pH resistant coatings, for delivering phage through acidic portions of the GI, including the stomach. Compositions may be therapeutic compositions formulated for oral administration. In the case of poultry and livestock, compositions may be delivered mixed or applied to animal feed or drinking water.

In some embodiments, the bacteriophage may be included as part of a composition for delivering the phage to the upper respiratory system, including the nasopharyngeal tract. Compositions include those suitable for delivery to the throat, nasal passages, sinuses, and ear (including ear canal or middle ear).

In another aspect, the present disclosure is also related to methods of preventing, ameliorating, treating, controlling, or preventing recurrence of a toxin-associated infection. The method comprises administering the bacteriophage described herein to a subject exhibiting symptoms of a bacterial toxicosis, or at risk of developing a bacterial toxicosis. A subject can be at risk of developing a bacterial toxicosis when exposed to other subjects having the condition, or exposed to spores that place the subject at risk, such as Clostridium spores. In some embodiments, the patient is hospitalized, and/or is immunocompromised, placing the patient at risk of nosocomial infection. In some embodiments, toxin-associated infection or toxicosis is associated with C. difficile infection (e.g., CD AD) or C. perfringens infection (e.g., necrotic enteritis).

In some embodiments, the present disclosure provides methods for treating Streptococcus pneumoniae infection, including pneumonia, otitis media, sinusitis, meningitis, and bacteremia. In some embodiments, the invention comprises administering the engineered phage against Streptococcus pneumoniae to a subject. In some embodiments, the subject may or may not receive antibiotic treatment concurrently. In some embodiments, the infection is antibiotic resistant.

In various embodiments, the engineered phage persists in vivo for at least several days (e.g., about 2 to 5 days, such as about 3, about 4, or about 5 days), or in some embodiments from 1 to 2 weeks. The persistence of the phage allows treatment or prevention of acute infection and toxicosis, without long term impact on the subject or microbiome of the subject.

For C. difficile infections, the current standard of care is treatment with gram- positive antibiotics. Approximately 20% of patients treated and cured of the acute infection experience recurrent infections, and treatment with antibiotics becomes less effective. The recurrence is thought to arise through the germination of spores of the bacterium in the gut, and this may be triggered by release of the toxins during sporulation, preparing the micronutrient environment in the gut to support germination by degrading the intestinal epithelium. Sequestration of the toxin prevents germination most likely (precise mechanisms remain unknown) by maintaining a nutrient poor environment for germination and allowing the gut epithelium to repair itself. Spores are eventually passed through the gut and eliminated in the feces. This is the basis for the passive immunization strategy with anti-toxin antibody therapeutics. The anti-toxin engineered bacteriophage described herein has distinct advantages over passive immunization. A lytic phage engineered against the pathogen itself would generate anti-toxin, thus reducing the likelihood of recurrence, and also kill the bacteria directly through its lytic activity. The alternative strategy of infecting commensal bacteria provides anti-toxin activity, but will not kill the target (toxin-producing) bacterium. This type of therapeutic approach may require concurrent use of standard-of-care antibiotics. In either case, targeted at C. difficile or a commensal bacterium, there is an advantage in continuous generation of the anti-toxin by the phage during infection, accumulating anti-toxin at the site of infection inside the gut. This has advantages over the passive immunization strategy, administered IV, whereupon the antibody must cross the gut wall to achieve efficacy and is not continually produced, but is a single bolus infusion.

For infections or prevalence of C. perfringens in poultry, the bacteriophage may be delivered to flocks through feed or drinking water. In these embodiments, the dose of the bacteriophage could be relatively low per bird, and thus could have significant economic and logistical advantages over flock vaccination.

EXAMPLES

Example 1 : Engineering of GRCS Phage against Staphylococcus aureus Staphylococcus aureus is a Gram-positive, catalase-positive, coagulase-positive, facultative aerobe that is an opportunistic pathogen with a high prevalence of antibiotic resistance genes and the ability to form biofilms on abiotic surfaces, rendering Sau infections often recalcitrant to standard antibiotic therapy. The development of phage therapy to treat Sau infections is therefore gaining prominence as an alternative to traditional antibiotic treatment.

Superantigens (SAgs) such as staphylococcal enterotoxins, carried by a number of chromosomal and mobile genetic loci, may contribute to the pathogenicity of Sau by helping to evade the host immune system through non-specific T-cell activation and cytokine release. Among the 22 known SAgs found in Sau, entertoxin B (SEB) is considered a Category B select agent by various United States federal agencies. SEB, when inhaled, can induce several symptoms within 120 min involving an aching feeling (head and muscles), increased heartbeat, coughing, enteric dysfunction (i.e., nausea, vomiting, and diarrhea), as well as eye irritation. Nanogram levels of inhaled SEB are incapacitating, while microgram levels can be fatal. While SEB is recognized as a potential biothreat due to inhalation, it is also known to cause about 10% of food poisoning outbreaks, whereas another SAg, staphylococcal enterotoxin A (SEA) accounts for approximately 80% of the cases of food poisoning outbreaks in the USA. Gastrointestinal disease caused by these toxins is usually self-resolving, and rarely lethal, although the elderly are more susceptible.

SEB is carried on a chromosomal Sau pathogenicity island (SaPI), a mobile genetic element that under stress conditions can excise from the chromosome, package into virus like particles and be transmitted to other Sau strains. SEB induces inflammatory cytokines, including TNF-a, interleukin (IL)-2, IL-6, IL-10, and interferon-γ, and chemokines, including monocyte chemoattractant protein 1, and normal T-cell expressed and secreted proteins. T cell-mediated toxicity is initially mediated by TNF-a induced inflammation, resulting in SEB-induced septic shock. Furthermore, SEB in vitro studies showed that SAgs activate TNF transcriptionally in T cells and monocytes, the major cause of shock syndrome. SEB alone can induce fatal disease, as demonstrated by intrapulmonary instillation of purified SEB inducing hemorrhagic lung tissue, respiratory distress, and lethal toxic shock syndrome in rabbits.

The three-dimensional structure of staphylococcal enterotoxin B has been determined to a resolution of 2.5 A. The structure contains an unusual two domain, main chain fold, a general motif adopted by most staphylococcal enterotoxins. A shallow cavity formed by both domains forms the T-cell receptor binding site, while the MHCII molecule binds to an adjacent site.

A strain, isolated from a human prosthetic joint infection, known to carry the SEB gene was tested in both planktonic and biofilm cultures for SEB expression and mitogenic activity. Transcription of SEB was significant and mitogenic activity was similarly high, suggesting that this strain produces biologically active SEB. This strain is also readily infected by the lytic dsDNA phage GRCS, as detected by the rapid increase in the major capsid protein gene expression by PCR upon infection by phage and rapid lysis in short term planktonic culture. Bacteriophage GRCS is a member of the Picovirinae subfamily of the Podoviridae family of viruses and has shown a high rate of infectivity toward Sau isolates that have been recovered from prosthetic joint infections.

To engineer an anti-toxin bearing phage, the natural GRCS phage genome was introduced into a bacterial artificial chromosome using PCR fragments from isolated phage DNA and reconstructed in the BAC vector using Gibson assembly. Transformation of the assembly mixture into E. coli amplified the construct, which was then isolated and used for transformation in the GRCS Sau host strain.

The GRCS/BAC transformation was allowed to recover for up to 3 hours and the recovery supernatant was tested for phage production by overlay plating on Sau strain RN4220. Plaque formation by natural phage transduction and the transformation supernatant were similar, illustrating efficient transduction by both natural phage and reconstructed engineered phage (Figure 1, left panel). Phage DNA from the respective plates was then prepared and amplified by PCR using primers used to prime the Gibson assembly fragment, and resulted in identical banding pattern on gel electrophoresis (Figure 1, right panel), demonstrating that plaques formed by the GRCS/BAC construct were GRCS phage.

With successful transformation of the GRCS-BAC construct and resulting phage generation, a reporter gene, green fluorescent protein, was inserted at two different loci in the GRCS genome by Gibson assembly in the BAC (Figure 2). The GFP gene is approximately 700 bp and encodes the 238 amino acid (26.9kDa) GFP protein. This is approximately the same size as a single-chain fragment variant antibody (scFv), with a molecular mass of approximately 27 kDa. The purpose of inserting GFP was to examine the different loci for gene expression from the phage genome for placement of an anti-toxin scFv or VHH antibody. We transformed the assemblies into E. coli, amplified and purified the constructs, followed by transformation into the Sau host strain (HFH-29994), and subsequent plating on the Sau host strain.

The GFP construct at loci 1 (GFP1) yielded transducing phage. Individual isolated plaques were tested for the presence of the GFP gene using the set of primers pictured schematically on the upper right (Figure 3). The loss of the GFP insert is clear, as the lane containing the BAC-GRCS-GFP1 control shows the distinct band at 1.3 kb, corresponding to the inserted GFP 1 gene. The plaques however, showed no presence whatsoever of the GFP1 gene. Detection of the phage DNA was confirmed using a set of primers (D and E) to another site in the GRCS genome. This suggested a loss of the GFP 1 gene upon transformation and recovery, and that insertion in this locus is unstable.

A similar observation was made with the second locus, GFP2. However, trace remnants of the GFP2 gene insertion can be seen in the transformation lysate after recovery (Figure 4). Passage of the recovery lysate to another liquid culture with fresh host cells resulted in complete loss of the GFP2 gene insert, however. Similar loss of the GFP2 gene insert was observed upon direct plating from the recovery lysate. Thus, insertion of GFP, at both loci, GFP 1 or GFP2, will require stabilization.

The unexpected presence of the GFP2 gene insertion in the recovery lysate, compared to insertion in locus 1, demonstrated some amount of stability at this insertion site. The extra DNA inserted at this locus in the phage genome could be destabilizing the phage virion. Packaged DNA internal pressure, due to charge density, is precisely balanced by stabilizing interactions among the head structural proteins of the phage virion, as counter-pressure in the natural phage. To rebalance the internal/external pressure, 0.5M sucrose was added to the transformation recovery. The recovery lysate + sucrose demonstrated increased recovery of the engineered GFP2- GRCS construct in the lysate (Figure 5). Passage in liquid culture, or plating on solid agar, in the absence of osmotic stabilization by sucrose, resulted in reversion to the natural phage.

To continue the engineered phage development for the anti-toxin phage, the insertion of the trans-genes will be stabilized by initially providing osmotic support to counteract the imbalance in head pressure. The phage head will be subjected to negative selective pressure for reversion to the natural phage by targeting a CRISPR- Cas endonuclease to natural phage revertant sequences at the site of trans-gene insertion, thus cutting the natural phage genome, while maintaining the engineered GFP phage genome integrity. The engineered phage containing the GFP trans-gene will be passaged for several generations under conditions of progressively decreasing osmotic support, in the presence of CRISPR-Cas negative selection for the natural phage. In this manner, the insertion of a 700 bp reporter gene will produce an intact phage particle that can accommodate an anti-toxin scFV or VHH antibody of approximately the same size. The evolved genome of the engineered phage will be used as a new scaffold for insertion of trans-genes of similar size for either measurement of expression using GFP at different loci, or insertion of the anti-toxin protein. In this particular example, the target for the anti-toxin protein would be SEB, and it would be expected that phage infection of an SEB carrying strain by the engineered anti-toxin phage would kill the bacteria by lysis, but also neutralize the cyto- and immune- toxic actions of SEB. Example 2: Engineering phage targeting Streptococcus pneumoniae for production of anti-pneumolysin antibody

Streptococcus pneumoniae (Spn) is a Gram-positive lancet shaped diplococci, and a common commensal organism found in the nasopharyngeal tract in human. Upon viral infection or other insults to the upper respiratory system, Spn can become pathogenic and cause pneumonia, otitis media, sinusitis, meningitis and bacteremia [1] . In the United States there are approximately 900,000 cases per year of pneumococcal pneumonia, of which 400,000 cases require hospitalization [2]. This is despite the availability of both polyvalent polysaccharide vaccine (PP23) and polyvalent conjugate vaccine (PCV 13, Prevnar) [3-6]. The high incidence of pneumococcal pneumonia is predominately among the adult population due to nearly 67 million adults not receiving adequate vaccination. In addition, the prevalence of drug-resistant Spn is high, with 30% of severe Spn cases resistant to one or more clinically relevant antibiotics, resulting in almost 1.2 million illnesses per year [7]. Importantly, vaccination has been shown to reduce the prevalence of antibiotic resistant Spn in the community [8]. The suboptimal adequate vaccine coverage combined with the high prevalence of antibiotic resistance requires the implementation of alternative treatment to combat pneumococcal disease, especially pneumococcal pneumonia.

An alternative treatment for pneumococcal pneumonia is engineered phage. Very few lytic phages that infect Spn have been identified. Phage Cp- 1 and SOCP are virulent members of the Picovirinae subfamily of the Podoviridae family of viruses that specifically infect Spn [9]. These phages have small genomes, ~19kB, with 345 bp inverted terminal repeats and covalently linked terminal proteins, required for DNA replication and packaging. These phage genomes have no significant sequence homology to the Sau phage GRCS (Example 1), but are organized in similar functional clusters, ordered differently, and transcribed in a single direction, left to right. The phage engineering methodology utilized for constructing the Sau phage GRCS would be amenable to engineering the Spn phages CP- 1 and SOCP.

A specific Spn toxin has been identified as playing a critical role in pneumococcal pneumonia, and especially the progression of pneumococcal pneumonia to bacteremia. Pneumolysin is a 52.7 kDa cholesterol-binding protein that oligomerizes in the eukaryotic membrane, forming a very large circular pore consisting of 42 membrane -inserted monomers, and roughly 350-450 A in diameter [10, 11]. In addition to this cytotoxic pore-forming activity in epithelial cells, pneumolysin also directly induces specific inflammatory responses at sub-lytic concentration [12-14]. Evidence of pneumolysin's role in infection has been demonstrated by the protective effect of antibodies to pneumolysin in attenuating aspects of pneumococcal pneumonia, including cytokine activation, inflammation, lung injury and infection [15, 16]. Inactive forms of pneumolysin have been used to generate protective antibody response and are broad spectrum, serotype independent, vaccine candidates [17, 18]. In addition, it has been shown that lytic antibiotics, such as β-lactams, can release significant amounts of pneumolysin from lysing pneumococcal bacteria, exacerbating potential lung injury in the treatment of pneumonia [19]. This concern has prompted discussion regarding the appropriate antibiotics for use in treating pneumococcal pneumonia, with a proposed preference towards non-lytic antibiotics such as macrolides and fluoroquinolones, especially in less severe forms of pneumonia to avoid unnecessary lung damage due to pneumolysin release.

Neutralizing pneumolysin as a payload would be advantageous in phage therapy. Natural phages with their lytic action would release pneumolysin from Spn, and thus may exacerbate lung injury in treating pneumococcal pneumonia, similar to the effect of lytic β-lactam antibiotics. The delivery of an anti-toxin molecule, via phage infection and bacterial expression, would neutralize the effect of pneumolysin from Spn, minimizing its inflammatory and cytotoxic impact on the lung epithelium. Thus, an engineered anti-toxin phage would provide superior efficacy in treatment of pneumococcal pneumonia, and other Spn related infections, by clearing the infection by Spn, but also ameliorating or preventing significant epithelial cell damage and inflammation due to pneumolysin. REFERENCES

1. Vernatter, J. and L.A. Pirofski, Current concepts in host-microbe interaction leading to pneumococcal pneumonia. Curr Opin Infect Dis, 2013. 26(3): p. 277, PMID: 23571695

2. Huang, S. S., et al., Healthcare utilization and cost of pneumococcal disease in the United States. Vaccine, 2011. 29(18): p. 3398, PMID: 21397721 Rodrigo, C, et al., Impact of infant 13-valent pneumococcal conjugate vaccine on serotypes in adult pneumonia. Eur Respir J, 2015. 45(6): p. 1632, PMID: 25792633

Tomczyk, S., et al., Use of 13-Valent Pneumococcal Conjugate Vaccine and 23- Valent Pneumococcal Polysaccharide Vaccine Among Adults Aged >65 Years: Recommendations of the Advisory Committee on Immunization Practices (ACIP). Morbidity and Mortality Weekly Report, 2014. 63(37): p. 822, PMID: Richter, S.S., et al., Pneumococcal serotypes before and after introduction of conjugate vaccines, United States, 1999-2011(1.). Emerg Infect Dis, 2013. 19(7): p. 1074, PMID: 23763847

Nuorti, J. and C. Whitney, Updated Recommendations for Prevention of Invasive Pneumococcal Disease Among Adults Using the 23-Valent Pneumococcal Polysaccharide Vaccine (PPSV23). Morbidity and Mortality Weekly Report, 2010. 59(34): p. 1102, PMID:

Office of Infectious Diseases and CDC, Antibiotic resistance threats in the United States, 2013. 2013, Centers for Disease Control and Prevention National Center for Emerging and Zoonotic Infectious Diseases (NCEZID) Division of Healthcare Quality Promotion (DHQP). p. 114.

Kim, L., et al., Biological and Epidemiological Features of Antibiotic-Resistant Streptococcus pneumoniae in Pre- and Post-Conjugate Vaccine Eras: a United States Perspective. Clin Microbiol Rev, 2016. 29(3): p. 525, PMID: 27076637 Leprohon, P., et al., A genomic approach to understand interactions between Streptococcus pneumoniae and its bacteriophages. BMC Genomics, 2015. 16: p. 972, PMID: 26582495

van Pee, K., et al., CryoEM structures of membrane pore and prepore complex reveal cytolytic mechanism of Pneumolysin. Elife, 2017. 6, PMID: 28323617 Solovyova, A.S., et al., The solution structure and oligomerization behavior of two bacterial toxins: pneumolysin and perfringolysin O. Biophys J, 2004. 87(1): p. 540, PMID: 15240487

Rai, P., et al., Pneumococcal Pneumolysin Induces DNA Damage and Cell Cycle Arrest. Sci Rep, 2016. 6: p. 22972, PMID: 27026501 13. Yoo, I.H., et al., Role of pneumococcal pneumolysin in the induction of an inflammatory response in human epithelial cells. FEMS Immunol Med Microbiol, 2010. 60(1): p. 28, PMID: 20528932

14. Witzenrath, M., et al., Role of pneumolysin for the development of acute lung injury in pneumococcal pneumonia. Crit Care Med, 2006. 34(7): p. 1947,

PMID: 16715037

15. Salha, D., et al., Neutralizing antibodies elicited by a novel detoxified pneumolysin derivative, PlyDl, provide protection against both pneumococcal infection and lung injury. Infect Immun, 2012. 80(6): p. 2212, PMID: 22473606 16. Cockeran, R., R. Anderson, and C. Feldman, Pneumolysin as a vaccine and drug target in the prevention and treatment of invasive pneumococcal disease. Arch Immunol Ther Exp (Warsz), 2005. 53(3): p. 189, PMID: 15995579

17. Mann, B., et al., Broadly protective protein-based pneumococcal vaccine composed of pneumolysin toxoid-CbpA peptide recombinant fusion protein. J Infect Dis, 2014. 209(7): p. 1116, PMID: 24041791

18. Lucas, R., et al., Mini-review: novel therapeutic strategies to blunt actions of pneumolysin in the lungs. Toxins (Basel), 2013. 5(7): p. 1244, PMID: 23860351

19. Brown, L.A., A.M. Mitchell, and T.J. Mitchell, Streptococcus pneumoniae and lytic antibiotic therapy: are we adding insult to injury during invasive pneumococcal disease and sepsis? J Med Microbiol, 2017, PMID: 28792379