Anthrax is a disease of animals including humans and results from infection by Bacillus (B.) anthracis 123. Anthrax spores reside in soil samples worldwide and are resistant to environmental insults such as temperature, moisture and UV irradiation. Spores enter host animals via inhalation, epidermal or gastrointestinal routes with respiratory route being the most fatal. Once inside hosts, the spores germinate and secrete three toxin components, called lethal factor (LF), edema factor (EF) and protective antigen (PA), encoded by the pXO1 plasmid. LF, a metalloprotease, plus PA is termed lethal toxin and EF plus PA is termed edema toxin 3. The combination of bacteremia and release of the protein toxins leads to sepsis, pulmonary edema and other fatal effects 12345.

PA is responsible for translocating the two other gene products, LF and EF, into the cytosol of susceptible cells 6789. The precursor form of PA (PA83) binds to ubiquitous cell surface receptors including von Willebrand factor, tumor endothelial marker 8 (TEM8) and capillary morphogenesis protein 2 1011. PA83 is cleaved by furin as well as by serum proteases 121314. The active form of PA (PA63) then heptamerizes and binds with a high affinity to LF or EF 3. The complex of PA with LF or EF forms a channel to allow LF/EF to translocate from the endosome to the cytosol where the toxic effects associated with LF are manifest 3. Cationic peptides that inhibit furin-mediated activation of PA83 to PA63 are also effective in blocking lethal toxin cytotoxicity 131617.

EF is a calmodulin-dependent adenylate cyclase and thus elevates intracellular cAMP levels of intoxicated cells 18. As a result of this mechanism, EF causes the additional pathological effects in the host although it is less virulent than LF. Recent studies have demonstrated that adefovir diphosphate is a potent inhibitor of anthrax EF, in vitro 19.

Anthrax LF is a representative member of the zinc-dependent endopeptidases family as demonstrated by the presence of the HEXXH zinc-binding consensus sequence 315. LF, an 89 kD protein, is one of the main virulence factors of anthrax 315. LF contains numerous anionic sites both within the active site and at distant sites 202122.

Macrophages are target cells of LF toxicity in animal model systems. Exposure of murine macrophages to lethal toxin resulted in rapid loss of cell viability 32324. Conversely, mice depleted of macrophages were not sensitive to lethal toxin 123. The mechanism of lethality has been attributed to release of cytokines or apoptosis as well as other mechanisms but is highly dependent upon the LF concentration 123526272829.

The mitogen activated protein kinase/extracellular signal-regulated protein kinase (MAPK/ERK) pathway is a major regulator for communication of extracellular signals to the nucleus and is involved in cellular adaptations to the environment 3031. In the cytosol, LF cleaves members of the mitogen activated protein kinase kinase (MAPKK) family in the N-terminal region including MAPKK family members 1–3 3132. The reduced levels of MAPKKs then prevent p38 kinase-mediated activation of immune mechanisms B. anthracis to evade host immunological mechanisms. Recently published studies have demonstrated that small molecules can inhibit LF activity and subsequently block LF-mediated cytotoxicity 212933343536.

At present, the only mechanism to fatally "intoxicate" cells with lethal factor is via host infection with B. anthracis spores that germinate in lymphatic tissues and secrete their toxin components. None of the three individual gene products of pXO1 are toxic in vivo 37. Inhibitors of other proteases such as angiotensin converting enzyme and HIV-1/HIV-2 proteases are effective and highly specific drugs for the treatment of chronic diseases 38 and therefore suggest a logical strategy for identifying anthrax lethal factor inhibitors. Based upon the demonstration of the anionic rich regions of LF 212229, we chose chemical libraries that included cationic compounds to test for LF inhibition. These studies were directed at identification of compounds that selectively inhibited LF both at the enzyme level then evaluation of their effects in cell culture based assays. An ideal therapeutic would penetrate susceptible cells and be effective protease inhibitors in "post-exposure" models of treatment.

In the initial phase of this study, we sought to identify compounds that selectively inhibited anthrax lethal factor enzyme activity. Such compounds were hypothesized to be potential lead molecules for optimization as drugs to treat B. anthracis infection. Since inhibitors of this protease were not known at the time, we chose to screen structurally diverse collections of individual compounds (a "library" consisting of ~500 compounds in several chemical classes) as one approach towards lead identification. We included simple linear cationic polyamines (n = 17) in the screening library with the hypothesis that they might bind to anionic sites on LF and thus block substrate cleavage. The data presented in this study show that spermine (a simple linear polyamine) is a concentration-dependent, sub-micromolar inhibitor of LF with reduced inhibitory potencies (termed selectivity) versus other bacterial and mammalian proteases. Polyamine analogs of spermine, including spermidine and ornithine were less active than spermine but still displayed concentration-dependent inhibitory effects as LF inhibitors.

Based upon literature demonstrating that both polyamines and aminoglycoside antibiotics bind to the N-methyl-D-aspartate receptor 4748, we also evaluated aminoglycoside antibiotics for LF inhibition. In our independent studies reported here and identified by other laboratories 495051 we found that gentamicin inhibited LF enzyme activity without inhibiting other proteases from bacterial and mammalian sources. We then showed that other compounds were more potent LF inhibitors than gentamicin. To further validate the mechanism, we tested cationic peptides (n~5) such as D- and L-hexaarginine as well as non-peptidyl cationic polymers including poly-L-arginine and poly-L-lysine (molecular weight ranges = 5,000–15,000); the larger cationic polymers (both peptidyl and non-peptidyl) were more potent inhibitors. While these large molecules will not be drug leads, they validated the mechanistic hypotheses of LF inhibition. Based upon these data, we concluded that neamine possessed the most relevant combination of drug-like properties and it was used as a scaffold for designing more potent and cell permeable analogs 52.

Aminoglycosides are effective antibiotics for the treatment of Gram-positive and Gram-negative infections as well as certain mycobacterial infections 5354. Their use, however, is limited by lack of oral absorption and toxicity at high doses including both ototoxicity and nephrotoxicity 55. Because of such toxicities, intravenous use of aminoglycosides in large and diverse age/health populations would pose a significant risk if used as prophylactic agents. Orally active/non-toxic compounds are still needed for protection against bioterrorist threats based on B. anthracis and its toxins.

High affinity polyamine interactions with nucleic acids are well known in both cell-free and cellular systems 5657585960. In anticancer studies, for example, exogenous polyamines are cytotoxic by depleting endogenous polyamine levels through feedback inhibition mechanisms 59. The cellular uptake of linear polyamines is well-recognized and numerous transporters have been shown to modulate polyamine levels within cells and organelles [60 and references therein].

We also sought to determine if the presence of DNA or RNA in the MAPKK cleavage assays would affect polyamine inhibition of LF activity. First, we demonstrated that nucleic acids did not inhibit LF activity at concentrations below those known to be present in human blood (< 80 μg/ml; Promega tech bulletin). Subsequently, we have also demonstrated that the presence of these nucleic acids prevent certain compounds such as neomycin but not neamine from inhibiting LF activity. One hypothesis we have considered is the role of size of the inhibitory molecules; smaller compounds such as spermine and neamine did not bind nucleic acids whereas larger molecules that inhibited LF (neomycin) were rendered less potent in the presence of 4–8 μg/ml of DNA or RNA. This is a unique pharmacological discovery first demonstrated in this effort.

All three chemical classes (linear polyamines, aminoglycosides and peptides) are highly charged molecules; they were not expected to be active in cell models of anthrax lethal factor cytotoxicity. This result was confirmed in our initial LF cytotoxicity studies with RAW 264.7 macrophage cells as all compounds were not active up to the highest concentration tested (500 μM). Recent studies 50, however, have demonstrated that aminoglycosides at "seemingly physiological conditions" inhibit LF and exhibit antibiotic activity against B. anthracis. In addition, linear polyamines enter cells by multiple mechanisms including active transport 485960. These different results highlight the need for continued research in this area including longer compound exposure periods.

Inhibition of the proprotein convertase, furin, by cationic peptides was expected since cationic hexapeptides and nonapeptides have been demonstrated to inhibit furin activity 121346. In contrast, however, the charged, nonpeptidyl compounds were inactive as furin inhibitors at concentrations up to 10 μM. It is likely that other small molecule cations will inhibit furin processing. Such compounds may be effective in a broad spectrum of diseases where furin cleavage of proteins play a role, such as Alzheimer's disease, viral infections and bacterial infections 13.

Taken together, the enzymologic and pharmacologic results presented in this study demonstrate the role of anionic sites on anthrax lethal factor that specifically bind cationic compounds from different chemical classes. Ultimately multi-drug and multi-dose combinations of antibiotics that suppress production of anthrax spores plus edema factor and LF inhibitors that target intracellular toxins will be employed to treat people exposed to anthrax gene products. Recently, an antiviral agent called adefovir was found to inhibit edema factor adenylate cyclase activity 19. Combinations of edema factor and lethal factor inhibitors along with Gram-positive antibiotics could be utilized to treat people exposed to anthrax or its gene products. As a spin-off of these efforts, researchers may want to determine if the presence/absence of nucleic acids affects the potency of selected compounds.

M.E.G. (principal investigator, chemical library composition, data evaluation, manuscript preparation), L.C. (execution of biochemical experiments and data analyses), D.N. (establishment of cell assays and execution of cell assays), O.S. (synthesis of compound in Table 1), S.O. (synthesis of compound in Table 1) and T.H. (target conception; critical intellectual discussion with principal investigator and manuscript evaluation/critique)