DE68905183T2 - METHODS AND COMPOSITIONS FOR THE TREATMENT OF INFECTIVES IN MAMMALS BY USE OF MEDICINAL PRODUCTS WITH A CONTENT OF HYMENOPTERIC POISON, PROTEIN-BASED OR POLYPEPTIDE-COMPONENTS THEREOF, OR ANALOGOUSLY SUCH POWERS. - Google Patents

Description

Introduction

The invention relates to the use of certain secondary agents derived from nature and also synthetic analogues thereof for improving the efficacy of other primary chemotherapeutic agents useful against bacterial, viral and cancerous infections and in particular for the efficacy of antibiotic agents. The identity of antibacterial, anticarcinogenic agents and in particular antibiotic agents and the activities and therapeutic use of these materials are well known. The secondary agents used in the invention for increasing the activity of these primary anti-infective agents are also known per se and have been used in some cases in medicine, but their ability to improve the activity of antibacterial, antiviral and anticarcinogenic agents and in particular antibiotic agents has not been previously recognized. Some of the secondary agents used in the invention are primarily obtained from the venom of species of the genus Hymenoptera, which include bees, bumblebees, wasps, hornets and fire ants.

Brief description of the invention

The invention is based on the discovery that the Hymenoptera venom, which consists of active proteinaceous or polypeptide components have been isolated from such animal poisons and analogues of such proteinaceous or polypeptide components increase or potentiate the activity of antibacterial, antiviral and anticarcinogenic agents and in particular antibiotic agents

The present invention arose from work described in a dissertation in veterinary science by Lorraine Smith Mulfinger entitled "Synergistic Activity of Honey Bee Venom with Antibiotics" submitted to the Graduate School Department of Veterinary Science of the Pennsylvania State University. References to earlier work by others have been abbreviated since full references are given in the bibliography of Mulfinger's dissertation and at the end of this application.

Background and state of the art

The use of antibacterial, antiviral, carcinostatic and anticarcinogenic substances, although widely known in the art, is still the subject of intensive ongoing research, much of which, in addition to the discovery of new active substances, is aimed at the discovery of means for improving the activity of known agents.

In fact, certain substances derived from bee venom have been studied and have been shown to be useful for certain specific pharmacological applications. For example, US Patent 4,444,753, issued on April 24, 1984, describes a composition containing a component obtained by deproteinization of an extract from a venom sac from bees. This product has an immuno-stimulating activity, a carcinostatic activity, an effect of increasing the antibacterial activity of an antibacterial substance and an effect of increasing the carcinostatic activity of a carcinostatic substance. The invention disclosed in this patent is directed to carcinostatic, immuno-stimulating and antibacterial agents containing the described composition. While this invention is similar in purpose to the present invention, it differs in that the bee extract is modified by its deproteinization so that it is negative in the biuret reaction and in the sulfosalicylic acid reaction.

US Patent 4,370,316, issued on January 25, 1993 to the same inventors as in the above-mentioned patent, also claims a method for treating a host animal having reduced immunity by administering an effective amount of the deproteinized extract from a bee venom sac.

Thus, while antibacterial, antiviral and anticarcinogenic substances are well known, it is also known that a deproteinized extract from the venom sac of a bee has certain useful activities, including antibacterial activity, antibacterial activity stimulating activity and immunostimulating activity, it was not previously recognized that proteinaceous hymenoptera venoms, proteinaceous or polypeptide extracts thereof and analogues of such proteinaceous or polypeptide components have a potentiating effect on substantially all antibacterial, antiviral carcinostatic and anticarcinogenic Such enhancers of the activity of such primary anti-infective agents not only increase the effectiveness of doses of such agents which are effective alone, but may also render effective small doses of such agents which were ineffective alone.

As noted above, the present invention relates to the use of hymenoptera venom, proteinaceous or polypeptide components thereof, and analogs of such polypeptide or polypeptide components to improve the activity of anti-infective therapeutic agents in general. However, to simplify the description of the invention, it will be discussed below only for the purpose of illustration in the use of honey bee venom or its proteinaceous extract melittin for improving the activity of antibiotics in the control of bacterial, viral and cancerous infections. Honey bee venom (HBV) was chosen because it is readily available. It is noted, however, that the venom of other hymenoptera and proteinaceous or polypeptide components thereof, as well as analogs thereof, are also useful in the invention to varying degrees. Similar anti-infective agents other than antibiotics can also be used according to the invention in the treatment of infections for which they have previously been used, but their effect is increased when used in combination with protein-containing hymenoptera agents.

As further background, it is pointed out that a variety of beneficial activities are attributed to honey bee venom. Some of these activities are scientifically documented, while others are based only on empirical data and the Folklore. The in vitro antibacterial activity of honey bee venom is well documented (Schmidt-Lange, 1951; Ortel and Markwardt, 1955; Fennel et al., 1968), but little effort has been made to put this activity to practical use. In the present invention, data from several empirical experiments indicated that the antibacterial activity of honey bee venom may have a significant in vivo effect in the presence of antibiotics. Based on these observations, experiments were prepared to study the interactions of honey bee venom with antibiotics using in vitro experiments, where the two components were evaluated without the contributing effects of the natural immune responses of the host animal.

In this study, three bacterial strains were first tested against three different antibiotics, using separate checkerboard titrations of honey bee venom with each antibiotic. Representatives of three major groups of antibiotics (penicillins, aminoglycosides, and polymyxins) were selected and tested to determine whether honey bee venom could enhance the antibacterial efficacy of the selected antibiotic. An antibiotic from a fourth major group was later studied as described below. Once synergy was demonstrated on a checkerboard, further consideration was performed using a simplified procedure. Two automated test plates for the minimum inhibitory concentration (MIC) test, where the susceptibility of eleven antibiotics was titrated simultaneously, were inoculated in parallel with bacterial cultures, with and without non-inhibitory doses of honey bee venom (HBV). Eight gram-positive and four Gram-negative organisms were tested using this system in an attempt to find classes of antibiotics each showing synergy with HBV and to determine the spectrum of synergistic activity of these combinations against different groups of bacteria.

In addition to testing the whole honey bee venom, the venom was fractionated by size exclusion chromatography. Each of the four fractions was tested to determine whether a specific component was responsible for the antibacterial activity and also acted synergistically in an antibacterial assay. The fraction containing melittin, previously identified as an antibacterial element of honey bee venom (Fennel et al., 1968), was shown to be active in its purified form and to be synergistically effective on a scale equal to that of the whole honey bee venom.

Furthermore, the activity of various analogues of the active components of Hymenoptera venoms was determined and compared with that of melittin.

Short description of the drawings

Figure 1 is a diagram of the amino acid sequence of nelittin;

Figure 2 is a graph of optical density versus hours after inoculation, showing the antibacterial activity of honey bee venom (HBV) on S. aureus;

Figure 3 is a graphical representation of optical density versus hours after inoculation with ampicillin and HBV against S. aureus;

Figure 4 is a graphical representation of optical density versus hours after inoculation with kanamycin and HBV against S. aureus;

Figure 5 is a graphical representation of optical density versus hours after inoculation with polymyxin B and HBV against S. aureus;

Figure 6 is a graphical representation of optical density versus hours after inoculation with ampicillin and KBV against E. coli;

Figure 7 is a graphical representation of optical density versus hours after inoculation with kanamycin and HBV against E. coli;

Figure 8 is a graphical representation of optical density versus hours after inoculation with kanamycin and HBV against E. coli;

Figure 9 is a graphical representation of optical density versus hours after inoculation with polymyxin B and HBV against E. coli;

Figure 10 is a graphical representation of optical density versus hours after inoculation with ampicillin and HBV against kanamycin-resistant S. aureus;

Figure 11 is a graphical representation of optical density versus hours after inoculation with kanamycin and HBV against kanamycin-resistant S. aureus;

Figure 12 is a graphical representation of optical density versus hours after inoculation with polymyxin B and HBV against kanamycin-resistant S. aureus;

Figure 13 shows the electrophoresis results of 50ug melittin protein;

Figure 14 is a graphical representation of optical density versus hours after inoculation showing the antibacterial activities of melittin/HBV against S. aureus;

Figure 15 is a graphical representation of optical density versus hours after inoculation with melittin/HBV and kanamycin against S. aureus;

Figure 16 is a graphical representation of optical density versus hours after inoculation with rifampicin and HBV against S. aureus;

Figure 17 is a graphical representation of optical density versus hours after inoculation with rifampicin and HBV against Ps. aeruginosa;

Figure 18 is a graphical representation of optical density versus hours after inoculation with polymyxin B and bumblebee venom against E. coli;

Figure 19 is a graphical representation of optical density versus hours after inoculation with polymyxin B and wasp venom against E. coli;

Figure 20 is a graphical representation of optical density versus hours after inoculation with polymyxin B and hornet venom against E. coli;

Figure 21 is a graphical representation of log 10 bacteria/ml blood versus treatment of a single treatment model of septicemia (polymyxin B and melittin interactions);

Figure 22 is a graphical representation of log 10 bacteria/ml blood versus treatment in repeated treatment models of septicemia (polymyxin B and melittin interactions);

Figure 23 is a graphical representation of optical density versus hours after inoculation with melittin/analogues and polymyxin B against E. coli;

Figure 24 is a plot of optical density versus hours post-inoculation, showing the relative potency of melittin and analogues;

Figure 25 is a graphical representation of the optical density versus hours after inoculation, showing the relative activities of melittin and analogues; and

Figure 26 is a plot of optical density versus hours after inoculation, showing the relative activities of nelittin and analogues.

Composition of poisons

Insect venoms are heterogeneous mixtures of biochemical compounds. Most venoms contain more than 90% protein. Toxins and enzymes make up this protein part and are the cause of direct cell damage. While many enzymes such as phospholipase A2, acid phosphatase and hyaluronidase are contained in most venoms, toxins and other biologically active peptides in animal venoms are highly species-specific.

Venom-producing insects all belong to the insect genus Hymenoptera. As in snake venoms, enzymatic activities such as phospholipase A2, hyaluronidase and acid phosphatase are present in all insect venoms. The toxin and peptide components, however, vary from species to species. (Tu, 1977b)

The venom of the Italian honey bee (Apis mellifera) is the most studied insect venom. The main component of honey bee venom is melittin. This peptide has a molecular weight of 2,847 daltons and comprises approximately 50% of the venom dry matter. A second peptide, apimin, is present in an amount of approximately 5% of the venom and various other peptides are present in trace amounts. (Hberman, 1972)

The venoms of other Hymenoptera contain peptides that have biological properties similar to those of melittin. Examples of Such peptides are Bombolitin IV from bumblebees, Megabombus pensylvanicus, Mastoporan from wasps, hornets and yellow hornets and Crabolin from European hornets. A common feature of these peptides is their amphiphilic nature. These peptides have been subjected to sequence analysis and their structures are well known. (A. Argiolas and JJ Pisano, 1985)

The antibacterial activity of honey bee venom

The bactericidal activity of honey bee venom was first documented in 1941 by W. Schmidt-Lange (1941). He tested E. coli and staphylococci and found that both were sensitive to honey bee venom. In addition, he noticed that the minimum inhibitory dose of honey bee venom for E. coli was much higher than that for staphylococci.

Not until another 10 years had passed, Brangi and Pavan (1951) investigated various extraction methods to isolate the antibacterial activity of honey bee venom. They found that the activity was present in both water and acetone extracts of the venom. They also showed that the activity was also stable on heating at 100ºC for 15 minutes.

In 1955, Ortel and Markwardt (1955) published the results of a study of the variability of the sensitivity of different bacteria to the antibacterial activity of honey bee venom. 196 bacterial strains were tested. The results showed that the tolerance of honey bee venom is much greater in gram-negative organisms than in gram-positive organisms. The ranges for bactericidal concentrations were given as 12.5 to 25 ug/ml for gram-positive bacteria and 1 to 10 mg/ml. The bactericidal activity is purified simultaneously with the red blood cells in the "direct hemolytic fraction". The name "melittin" has not yet been assigned to the active component of this fraction.

In 1963, Benton et al. published a bioassay for honey bee venom. The bacteriostatic activity of the venom was quantitatively determined by a radial diffusion test in which zones of growth inhibition were measured caused by serial venom dilutions in a bacterial growth lawn. These tests were proposed to standardize the biological activity of honey bee venom intended for in vivo use. (Currently, allergy desensitization is the only in vivo honey bee venom treatment approved by the United States Food and Drug Administration.) The article also tested the thermal stability of honey bee venom activity and found that it could withstand sterilization procedures (121ºC for 15 minutes) (Benton et al. 1963)

Melittin isolation and activities

Honey bee venom contains several pharmacologically active compounds. The compound that is the most abundant in the venom is melittin, a polypeptide with a molecular weight of 2,847 daltons that acts directly as a hemolysin of red blood cells. Other active components include phospholipase A2, histamine, dopamine, noradrenaline, apaamine, and hyaluronidase (Haberman, 1972).

Antibacterial activity of melittin

Fennel, Shipman and Cole (1968) purified melittin by Sephadex G-50 chromatography and demonstrated that the melittin fraction possessed "potent antibacterial activity". They tested 30 random bacterial strains (including various streptococci, staphylococci and enterobacteria strains) comparing the activity of purified melittin with whole bee venom. They found that one strain of S. aureus, a penicillin-resistant isolate, showed no reduction in susceptibility to melittin.

Although melittin has been described as an antibacterial factor of honey bee venom, no reports of its in vivo activity have been found. It was noted by Mollay and Kreil (1974) that interactions between melittin and lecithin increased the phospholipase activity of honey bee venom on lecithin. However, it had not been previously recognized that melittin increases the activity of antibiotics.

Haberman and Jentsch (1967) purified melittin and published the amino acid sequence. They found that melittin exists in two natural forms, which differ only by a formyl substitution at the N-terminus (Figure 1).

Analogues of protein and polypeptide components of Hymenoptera venoms

The following analogues of melittin were prepared. No. of analogue compositions Melittin Mastoporan

The analogues were prepared by conventional peptide synthesis, as described, for example, by M. Bodanszky: "Principles of Peptide Synthesis", Springer Verlag, 1984.

As an example, the peptide synthesis of analogue No. 4, namely melittin (1-20)-Lys-Arg-Lys-Arg-Gly-Gly-NH₂, is now described in detail.

A derivatized resin, such as a polydimethylacrylamide gel, which is commercially available under the trade name PEPSYN KA, is treated with (Fmocly)₂O, where Fmoc is 9-fluorenylmethoxycarbonyl, which serves as a temporary protecting group. The reaction, which is carried out in the presence of 4-dimethylaminopyridine as a catalyst, leads to the formation of the ester Fmoc-Gly-O-resin.

The ester is deblocked in the presence of 20% piperidine in DMF to form H-Gly-O-resin.

The deblocked product is then treated with an activated ester of the formula

Fmoc-Gly-OPfp

activated, where Pfp is pentafluorophenyl, where

Fmoc-Gly-Gly-O-Resin

is formed.

The remaining 24 amino acids are coupled with the reaction product, which is formed in 24 similar deblocking and coupling cycles with active esters.

The product thus formed is deblocked with 20% piperidine in DMF and the melittin formed is cleaved from the resin in the presence of TFA (trifluoroacetic acid) and a rinsing agent such as water.

Antibiotics

Antibiotics can be functionally divided into four groups based on the active sites of the antibiotics (Volk, 1978a). Target structures of the four groups are the cell wall, the cell membrane, and the protein synthesis apparatus and the nucleic acid replication apparatus. Due to the complexity of the synergy test, four antibiotics were selected for testing, one from each of the aforementioned groups. The antibiotics selected were ampicillin, kanamycin, polymyxin B and rifampicin. Each has a different type of action on procaryotic cells.

Ampicillin

Ampicillin belongs to the group of antibiotics that affect the cell wall structure. This Antibiotics are all penicillin derivatives, each containing the ß-lactam functional ring. Collectively known as the β-lactam group, these antibiotics block cell wall synthesis by inhibiting the transpeptidase enzyme that cross-links the pentaglycine bridges of peptidoglycan, with their presence affecting only actively growing cells.

Ampicillin is a semi-synthetic derivative of penicillin. The synthetic step in ampicillin synthesis adds an amino group to the α-carbon of penicillin G. This confers β-lactamase resistance (the predominant penicillin resistance factor of bacteria) which gives ampicillin a wider spectrum of activity among bacteria than penicillin (Volk, 1978b).

Kanamycin

Kanamycin is an aminoglycoside. This group of antibiotics blocks protein synthesis. Members of this group find the 30 ribosomes of bacteria and steric blocks to bind aminoacyl-tRNA or inhibit the relocation of the growing peptide chain to the active ribosomal center (Volk, 1979c). Since protein synthesis is required for many regulatory cell functions, aminoglycosides are effective for bacteria in any active or stationary phase of growth

Polymyxin B

Polymyxin B is a cyclic aliphatic peptide. Due to its combined hydrophilic and hydrophobic properties, polymyxin B has a detergent-like action that does not require cell growth to be effective. As with melittin, polymyxin B interacts with the membranes to form small hydrophilic pores in the hydrophobic regions of the membranes. In gram-negative organisms, which have a thick lipopolysaccharide layer that acts as a selective permeable barrier, polymyxin B is effective in destroying osmotic gradients. Consequently, polymyxin B is very effective against gram-negative organisms, while it is only minimally effective against gram-positive organisms (Sebek, 1979). While melittin can form membrane pores similar to polymyxin B, melittin is more active against gram-positive organisms, so the action of melittin cannot be completely analogous to that of polymyxin B.

Rifampicin

Rifampicin is an antibiotic from the group acting at the stage of nucleic acid synthesis, which completes the examples of antibiotics from the four main groups referred to above.

Synergy studies

A review of articles examining synergy between antibiotics and other compounds in bacterial systems shows that all researchers used essentially the same basic approach. Bacterial growth was monitored in broth cultures with and without each compound separately and then with both compounds together. To prove synergistic action as opposed to an additive effect, at least one of the compounds used would be shown to be effective at the stage where it used alone, show minimal growth inhibition. With a relatively inactive compound, therefore, any increased activity of the second compound in its presence would be the result of synergistic interactions (Moellering et al., 1971; Carrizosa and Levison, 1981; and Cynamon and Palmer, 1983). It is on this type of arrangement that the experiments of the present invention were based.

Materials and methods materials

Honey bee (Apis melifera) venom was supplied by Vespa Laboratories, Spring Mills, Pennsylvania.

Bacterial strains were provided by the Department of Veterinary Sciences at Pennsylvania State University. S. aureus #140A is a field isolate from a case of bovine mastitis. E. coli #G1880E was selected from the systematic collection of the E. coli Reference Center. A kanamycin-resistant strain of S. aureus was isolated by a natural selection process as described below.

Antibiotics were purchased from Sigma Chemical Company (St. Louis, Missouri) and the activity units were based on their analyses.

Tryptic soy bases (BBL Microbiology Systems, Cockeysville, Maryland) were used to support bacterial growth both in broth and on agar.

Sephadex R G-50 was purchased from Pharmacia Fine Chemicals, Uppsala, Sweden.

Minimum inhibitory concentration (MIC) tests of Antibiotics with and without honey bee venom were tested by the Department of Microbiology, Allegheny General Hospital, Pittsburgh, Pennsylvania, using the Sensititre TM test system distributed by Gibco Laboratories, Lawrence, Massachusetts.

Proceedings Isolation of kanamycin-resistant mutants

S. aureus was grown in 5 ml of tryptic soy broth (TSB) overnight at an approximate density of 10⁹ colony forming units/ml. 0.1 ml of the overnight cultures was plated on a plate of trypticase soy agar (TSA) containing 39 ug/kanamycin and incubated for 48 hours at 37°C. Colonies that appeared within 48 hours were subcultured on a second TSA plate supplemented with 39 ug/kanamycin.

Checkerboard titration test for synergy

Bacterial cultures were prepared for this assay by freezing each strain during logarithmic growth in TSB. For this purpose, a 5 ml overnight culture was used to inoculate 200 ml of TSB in a 500 ml Erlenmeyer flask. The culture was incubated at 37ºC with constant agitation and the optical density (OD) at 600 nm was read hourly. When the culture reached mid-log phase (approximately at 0.55 OD units), 5 ml aliquots were transferred to 16 x 100 mm screw cap tubes. All cultures were frozen and stored at -20ºC. E. coli required glycerol addition to the medium to a final concentration of 20°C to survive freezing. This was achieved by mixing 1 ml of sterile glycerol with 4 ml of the log phase culture immediately prior to freezing. To begin an assay, a tube of frozen culture was thawed in a beaker of water at room temperature. The thawed culture was added to 175 ml of TSB in a 500 ml Erlenmeyer flask, stirred, and the OD660 immediately measured and recorded as "time zero." The flask was then incubated at 37°C with constant stirring for 2 hours, after which time the OD660 was again measured and recorded, the culture was divided into 16 x 100 mm screw-cap tubes which were prepared by adding the specific aliquots of honey bee venom (HBV) and the antibiotic described below.

Stock solutions of HBV and antibiotics were prepared in distilled water, filter sterilized and stored at -20ºC in 5 ml aliquots at concentrations twice the concentration required for the checkerboard titration system. The frozen stock concentrations required for each bacterial species are given in Table 1. The concentration used for each bacterium was based on preliminary experiments in which the antibiotics were used alone to determine the minimum inhibition ranges of each antibiotic for each microorganism.

For each test, one ampoule of antibiotic and one ampoule of HBV were thawed at room temperature and diluted with an equivalent volume of 2 x TSB and then serially diluted twice into a normal strength DSB to obtain four poison concentrations and four antibiotic concentrations. Seventy-five screw-cap test tubes were numbered and arranged to conform to the checkerboard pattern shown in Table 2. TSB, antibiotic and HBV were then distributed according to the arrangement shown in Table 3. Tubes designated OO and O contained 2.5 ml of TSB and served as OD blanks and sterility control tubes. Tubes 1-75, each containing a total volume of 5 µl, were inoculated with 2 ml of the two-hour culture described above [Note: The final HBV and/or antibiotic concentration in each tube was one-tenth of the concentration added to the 200 µl aliquot (see Table 3).] Each tube was immediately capped and inverted. After all tubes were incubated, they were placed in horizontal racks placed on a rocking platform at 37ºC. The growth of each tube was monitored individually at 4, 6, 8, 12 and 24 hours by determining the optical density of each tube at 660 nm.

The minimum inhibitory concentration tests with HBV

The Allegheny General Hospital microbiology laboratory, which has the capability to perform automated MIC testing, was contacted to conduct a test trial on 12 clinical bacterial isolates. The adaptation of the automated MIC test included the following limitations: (1) each trial could only test one dose level of HBV, and (2) the effect of HBV alone could only be determined as inhibitory or non-inhibitory. The synergy of HBV with 11 antibiotics in this system was evaluated by conducting two trials performed simultaneously with and without HBV present. The HBV dose used for each species was assumed to be a non-inhibitory dose based on the checkerboard titration test.

Melittin Cleansing

Sephadex R G-50 gel for gel filtration was swollen for 24 hours at room temperature in β-alanine acetic acid buffer (BAAB), pH 4.3 (Guralnick et alia, 1986) and then equilibrated at five degrees Celsius overnight. It was loaded into a 2.5 x 60 cm column and equilibrated at a flow rate of 1.0 ml/h. 100 mg of HBV was reconstituted in 5 ml of BAAB buffer containing 20% sucrose. The HBV was applied to the column and eluted at a flow rate of 1 ml/h. The eluates were monitored at an absorbance of 280 nm. The fractions containing the main peak were pooled, an aliquot was tested by the Lowry protein test (Lowry, 1951) and the remainder was freeze-dried.

Identification of the melittin fraction was based on relative mobility and quantification of bands that appeared in polyacrylamide gel separations for each fraction (Benton, 1965). Melittin was also tested for purity by polyacrylamide gel electrophoresis. Electrophoresis was performed as described by Guralnick et al. (1986). Freeze-dried fractions were reconstituted in two mg/ml electrophoresis sample buffer and 50 μl samples were applied to the gel per sample well.

Full-toxic equivalence of melittin

The amount of the equivalent melittin fraction versus its content in whole honey bee venom was determined by quantification of individual bands in electrophoresed samples of whole venom and the melittin fraction. 20, 40, 60, 80 and 100 µg samples of whole honey bee venom were separated by electrophoresis, stained with Coomassie brilliant blue perchloric acid stain and scanned with a densitometer. A standard curve was constructed in which the peak area of the melittin band of the whole venom samples was compared to the amount of protein in the applied sample. Six 40 µg samples of the purified melittin were tested simultaneously and their equivalence in honey bee venom was tested using the standard curve. This procedure is described in detail by Mulfinger et al. (1986).

Testing of melittin for synergistic activity

To compare the antibacterial activity of whole honey bee venom and the melittin fraction, the previous checkerboard titration results were reviewed and the test system was chosen such that HBV dose effects could be easily observed alone and in combination with an antibiotic. Since staphylococci are sensitive to HBV alone at the concentrations used in the above checkerboard tests, and since kanamycin showed a good synergistic effect with HBV, this system was chosen to compare the antibacterial activities of whole HBV and melittin. The dose of each component used in this analysis was 2 ug/ml HBV and 2.5 ug/ml kanamycin. These doses were in a range of bacterial reactivity in which the Effects of small dose changes were reproducible and easily measurable. The equivalent dose of melittin fraction for 2.0 ug/ml HBV was 1.6 ug/ml. Each experiment was tested in parallel triplicate samples of melittin fraction and whole honey bee venom and without kanamycin presence to check for equivalent activity.

Statistical analysis

Each checkerboard experiment was repeated five times. The means of the five replicates for each bacteria-antibiotic combination were tested for significant differences at each time point using a Waller-Duncan K-ratio T test and the families of curves were selected for the synergy test. One family of curves consisted of an experimental control curve (bacterial growth without antibiotic or HBV presence), an antibiotic control curve (bacterial growth with antibiotic but without HBV presence), a toxic control curve (bacterial growth with HBV but without antibiotic presence) and an interaction curve (growth with antibiotic and HBV presence). Families in which the antibiotic control curves and the poison control curves showed small mean OD decreases with respect to the test control curve, and which showed equally large OD decreases in the interaction curve with respect to the test control curve, were tested for synergy.

A synergistic effect between compounds can be differentiated from an additive effect of the compounds because the additive effect is predictable. Additive effects can be predicted by considering the individual effects of the two compounds, with any greater effect indicating a synergistic interaction. An equation was constructed that could predict the OD readings for an additive interaction between HBV and an antibiotic. See the Mulfinger dissertation (1987) cited above, pages 23-25.

Results Checkerboard titration experiments

Three bacterial strains were tested against each of the three antibiotics in combination with honey bee venom. These nine combinations of bacteria, antibiotics and HBV were analyzed using the checkerboard assay, with 25 treatments (antibiotic and HBV combinations) for each bacteria-antibiotic combination. Each checkerboard experiment included triplicate samples for each treatment and was repeated five times. The data from triplicate samples repeated in five experiments were averaged and the mean and standard deviation for each time point of each treatment is given in the appendix. For each bacteria-antibiotic combination, the mean OD values for each antibiotic/HBV treatment at each time point were arranged in descending order and grouped according to significant differences using the Waller-Duncan K-ratio T test. From the Waller-Duncan profiles, families of four curves were compared as described in the "Statistical Analyses" to demonstrate synergy. The curve families showed the largest OD differences between interaction curves and the lowest of the test curves, the antibiotic control curve and the poison control curve were plotted for each test point and tested for synergy using the equation mentioned in the "statistical analysis" section above. For each family of curves, if the estimate of (-X+A+V-AV) for a time point is significantly greater than zero, at 95% confidence level (ie, when synergy is indicated), the time point on the interaction curve is indicated by a superscript "s" next to the square representing that time point (Figures 2-11).

S. aureus

S. aureus is sensitive to honeybee venom alone at low concentrations. It is therefore important to find the maximum honeybee venom dose for which no effects are exerted. This concentration was approximately 2 ug/ml. Thus, for all antibiotic/HBV combinations with S. aureus, the venom doses for the checkerboard titration system were 0, 2, 4, 8, and 16 ug/ml (Tables A-1 to A-3). Figure 2 demonstrates the effects of these doses of honeybee venom when used alone as an antibacterial compound.

S. aureus against ampicillin/HBV

The final concentrations of ampicillin in the checkerboard tubes were 0, 0.05, 0.1, 0.2 and 0.4 ug/ml. Figure 3 shows the results of the ampicillin/HBV combination using 2 ug/ml HBV and 0.05 ug/ml ampicillin. No synergy was observed at the 4 or 6 hour time points; at the 8 and 12 hour time points, However, it is obvious that the interaction curve is much smaller than could be predicted from the sum of the effects caused by ampicillin and HBV alone. Statistical analysis shows that at both time points the summation (-X+A+V-AV) is significantly greater than zero.

S. aureus against kanamycin/HBV

The final concentrations of kanamycin chosen for testing S. aureus in the checkerboard system were 0, 1.25, 2.50, 5.0, and 10.0 ug/ml (Table A-2). Figure 4 shows the family of curves showing the greatest contrast between the control and interaction curves. In the experiment, the synergy is first demonstrated near 6 hours and is clearly visible at 8 hours of incubation. At 12 hours, the cultures appear to be outside the effect of the combined dose and the synergistic effect is lost because growth is limited by other factors (nutrients) in the medium. (This growth limitation is shown by the control curve). Despite the 12-hour growth limitation, statistical analysis of the data at 6, 8, and 12 hours demonstrates a synergistic interaction between kanamycin and HBV in this assay.

S. aureus against polymixin B/HBV.

The final concentrations of polymyxin B in these experiments were 0, 313, 624, 1250 and 2500 U/ml (Table A). Synergy was observed at 4 ug/HBV and 625 U/ml polymyxin B (Figure 5). Synergy is demonstrated by the interaction curve at both 8 and 12 hours of incubation.

E. coli.

The honey bee venom alone was not inhibitory against E. coli in quantities required for the demonstration of synergy (Tables A4-A6), thus toxicity was not the limiting factor for HBV in the checkerboard assay with E. coli. However, experimental conditions limited the upper HBV concentration to approximately 40 ug/ml; concentrations greater than this caused precipitation of the medium components. Consequently, the final HBV concentrations used in the checkerboard assays with E. coli were 0, 5, 10, 20 and 40 ug/ml

E. coli against ampicillin/HBV.

The final ampicillin concentrations used for use in the E. coli checkerboard titration were 0.5, 1, 2, and 4 ug/ml (Table A-4). The synergy was less dramatic in all families of curves evaluated than in any of the above experiments. Evidence of synergy was only present in the 40 ug/ml HBV-1 ug/ml ampicillin combination and only at the 6 hour time point (Figure 6).

E. coli against kanamycin/HBV.

The final kanamycin concentrations selected for the checkerboard assay were 0, 5, 10, 20, and 40 ug/ml (Table A-5). Figure 7 shows the effects of honey bee venom with a minimal effective dose of kanamycin. In this situation, only the 8-hour time point shows synergy. Regardless of the HBV dose, no synergy was observed in other combinations of HBV with lower doses of kanamycin.

Figure 8 shows a higher dose of kanamycin with HBV on E. coli. Here, synergy is statistically proven at all time points after two hours.

E. coli against polymyxin B/HBV.

The final concentrations of polymyxin B in the checkerboard titrations were 0, 1.5, 3, 6 and 12 U/ml (Table A-6). The combination of 3 U/ml polymyxin B and 5 ug/ml HBV provided the most dramatic illustration of synergy (Figure 9). Synergy is evident at all points during treatment and the differences between observed and predicted values are large.

Kanamycin-resistant S. aureus

A kanamycin-resistant S. aureus, obtained by selection of spontaneous mutants, was tested to evaluate the effect of HBV on drug-resistant bacteria. A kanamycin-resistant S. aureus was desirable because some synergy can be observed in this organism for all antibiotics and because the synergistic effects are most easily seen with kanamycin.

No differences in susceptibility to HBV were found between the resistant strains, with the toxin concentrations in the checkerboard experiments being the same as for the parental strain, 0, 2, 4, 8 and 16 ug/ml (Tables A-7 to A-9). It was noted that under identical conditions the resistant strains had a slower growth rate than the parental strains, thus a comparison of optical densities between experiments on two different strains is not meaningful.

Kanamycin-resistant S. aureus against ampicillin/HBV.

The final concentrations of ampicillin used in this checkerboard experiment were the same as for the parent S. aureus, 0, 0.05, 0.1, 0.2 and 0.4 ug/ml (Table A-7). Due to either the slower growth rate or the resistance factor, the effects observed with this strain were not completely analogous to those of the parent strain. The best Evidence of synergy was observed at higher ampicillin concentrations than for the parent strains. Because of the slower growth rate, a longer growth period was considered. Figure 10 shows the interaction of 2 ug/ml HBV and 0.4 ug/ml ampicillin. Statistical analysis of the data shows synergy at 8, 12 and 24 hours.

Kanamycin-resistant S. aureus against Kanamycin/HBV

The dosage of kanamycin required to reduce the growth rate of kanamycin-resistant strains of S. aureus was approximately four times the dose required for the original strain. The checkerboard experiment was within the range of kanamycin-resistant S. aureus and was 0, 5, 10, 20, and 40 ug/ml kanamycin (Table A-8). Again, the slow growth rate required a longer growth period to be considered. The combination of 8 ug/ml honey bee venom and 10 ug/ml kanamycin is shown in Figure 11. Although the dose of kanamycin used is twice the dose required for the original S. aureus, it remains effective twice as long in the presence of honey bee venom. Synergy was only observed after 12 hours and was only shown to be significant at 24 hours.

Kanamycin-resistant S. aureus against Polymyxin B/HBV.

It was interesting to note that this mutant, which was selected for enhanced resistance to kanamycin, became more sensitive to polymyxin B than the parent strain. Polymyxin B doses used for the checkerboard assay were 0, 12.5, 25, 50 and 100 U/ml (Table A-9), while the polymyxin B dose range used to test the original strain was between 312 and 2,500 U/ml. Figure 12 shows kanamycin-resistant S. aureus against 50 U/ml polymyxin B and 4 ug/ml HBV. Synergy was demonstrated at time 12.

MIC tests of antibiotics with and without HBV

The results of a preliminary study of the effect of HBV on the MIC of antibiotics against 8 gram-positive bacteria and 4 gram-negative bacteria are shown in Table 4 and Table 5, respectively. Despite the obvious inadequacies of the test system, definite trends could be observed in the results of the survey experiments. There was a strong assumption of synergy, where observations within a single MIC experiment showed that identical doses of HBV affected some MICs while not affecting others. In Tables 4 and 5, a (+) was used to indicate a decrease of more than a two-fold dilution of the MIC of an antibiotic in the presence of HBV. A (-) indicates no difference or only a single dilution step variation (judged as a variation of the test) in the MIC in the presence of HBV.

Table 4 shows the results from various gram-positive organisms. The results indicate that trends were present within the species tested. For example, S. aureus appears to show synergy with all antibiotic/HBV combinations, while S. epidermidis showed consistent synergistic results only with the cephalothin/HBV combination and sporadic results with other antibiotic/HBV combinations. The one Streptococcus faecalis strain tested did not reflect the same synergistic trends shown by the two staphylococcal organisms. The data in Table 5 lists the results of four E. coli strains in the MIC test system. Definite patterns of synergy were observed with each of the β-lactam antibiotics (ampicillin, carbenicillin and piperacillin) in the MIC test system. Also, the MIC of the amino glycosides gentimicin and amikacin were reduced in all cases except one. The MIC of cefoxitin was also reduced by HBV in all E. coli tests.

Melittin purification and testing Chromatography of honey bee venom

Purification of melittin on Sephadex G-50 gave well-defined, baseline-resolved peaks. The void volume was 100 ml and the eluted melittin fraction was between 100 and 230 ml after the void volume. Approximately 65 mg of the original 100 ml sample was found in fractions 200 to 230. These fractions were pooled and tested for purity by polyacrylamide gel electrophoresis. Figure 13 shows the electrophoretic results of 100 ug protein from pooled fractions 200 to 230. Comparison of the relative mobility of these bands to the relative mobilities of the electrophoretically separated HBV components identified melittin as the only component of fractions 200-230 that was detectable in this separation.

Testing of melittin for antibacterial activity

Equivalent doses of melittin and whole honey bee venom were tested for antibacterial activity in combination with and without antibiotic (Table A-10). Since S. aureu was sensitive to HBV at the concentrations used in the above tests, this organism was chosen to to test the activity of the melittin fractions. Kanamycin was chosen to evaluate the synergistic activity of the fraction because the interaction curve observed in the above testing of S. aureus against this antibiotic with HBV reflected HBV synergy at all time points.

Antibacterial activity of melittin.

The results of melittin against whole HBV are shown in Table 14. No significant differences were observed in the antibacterial activity of whole HBV and the melittin fraction. For each time point shown in Figure 14, the optical densities of the HBV curve and the melittin curve are statistically equal.

The synergistic activity of melittin with kanamycin.

Figure 15 compares the antibacterial activities of equivalent doses of the melittin fraction and whole HBV venom in combination with equal doses of kanamycin. None of the optical densities at any time point on the two curves are significantly different. Moreover, ignoring statistical evaluations, the interaction curve representing the melittin fraction is actually slightly lower than the interaction curve representing HBV at all time points. Thus, if the time points on both curves were taken as true means, the conclusion would be that the melittin fraction is indeed more active than whole HBV.

Interpretation of the chessboard test results

The results of the checkerboard tests clearly demonstrate a synergy between antibiotics and honey bee venom. Figure 2 illustrates the effects of different doses of honey bee venom on S. aureus without antibiotics. It can be seen in this figure It can be seen that the addition of high doses of venom, such as 8 or 16 ug/ml, to the growing cultures actually reduced the optical density of the culture. This indication of cell lysis is evidence that honey bee venom is indeed bactericidal. The mechanisms of this bactericidal activity and its contribution to the observed synergy with antibiotics are not known. The various results of the checkerboard titration tests indicate that several different synergistic mechanisms may be operating in these experiments.

Questions may be raised regarding the large standard deviations observed at different time points in the data tables. The variability is generally due to the sharp increase in growth rate when the bacteria are in log phase. Time points taken in the mid-log phase have a much larger difference in optical density versus time than time points taken during a slower growth period. Consequently, uncontrollable small variations in sampling intervals can cause larger variations in optical density readings at time points during logarithmic growth. As cultures are divided into different treatment groups during log phase, the variation between experiments is even more pronounced. However, this type of error is taken into account in the statistical analysis procedure. When using a large number (15), the estimation ranges for the means of the time points were made narrow enough to statistically evaluate the differences of these means.

Although melittin was originally only developed as a synergistic Component of HBV in combination with an antibiotic was tested with a bacterial strain for the purpose of discussing possible mechanisms, it must be assumed that melittin is the synergistic honey bee component in each of the bacterial antibiotic-HBV combinations tested.

Apparent increasing dosage

In most cases, honey bee venom appears to enhance the initial efficacy of the antibiotic, as indicated by an increasing ability to reduce bacterial growth levels immediately after addition of the two components. This type of synergy was best demonstrated with E. coli against HBV and polymyxin B (Figure 9). Synergy is evident at the first time point after addition and continues until the culture has passed log phase. These results lead to the conclusion that low doses of ineffective antibiotics can be made effective by the addition of HBV.

The dosage enhancement effect described above is the type of synergy seen in most of the experimental combinations tested. This type of effect could be explained by the action of melittin through several different mechanisms: (1) altering the solubility properties of the antibiotic molecules, (2) increasing the permeability of the bacterial membrane, and (3) increasing the potency of the antibiotic molecules at their active sites.

Changed solubility properties of the antibiotic.

Melittin could increase antibiotic efficacy by being transported more easily into the bacterial cell. The direct interaction of melittin with antibiotic molecules, which makes the molecules less polar or more hydrophobic, could allow passive transport across bacterial membranes. The amphiphatic nature and basicity of melittin makes it a possible candidate for such a function and adds to the plausibility of this mechanism. This type of mechanism could be similar to the one that facilitates the diffusion of potassium ions with valinomycin.

Improved membrane permeability.

The apparent dosage of an antibiotic could also be increased by reducing the permeability barriers of the bacteria.

Although this role as a pathway-forming peptide could easily be supported, it cannot be the only function of melittin involved in the antibacterial synergy. The increased transport across membranes cannot explain why melittin alone is more effective against gram-positive organisms that have a compromised membrane barrier.

Improved antibiotic specific activity.

A third possible mechanism for the synergistic interactions proposes direct interaction of melittin and the antibiotic to make the antibiotic more effective once it has reached the active site. A more specific example is a possible interaction with kanamycin. Once kanamycin has reached the 30S ribosome, a melittin-kanamycin complex may have a higher affinity for the active site than the unfound kanamycin (after all, melittin is a basic molecule like nucleic acids) or the melittin-kanamycin complex may be more effective at sterically blocking the ribosome's transfer RNA simply because of the size of the complex.

Increased active life of antibiotics

In several cases it was difficult to demonstrate an increase in antibiotic efficacy by the addition of honey bee venom (melittin) until late in the growth period. In these cases it appeared that melittin caused an increase in the efficacy of the antibiotic. This effect was not seen with kanamycin-resistant S. aureus treated with kanamycin/HBV. A relatively high dose of HBV is shown in Figure 9, the reasons for this being that no synergy was seen at lower doses. Thus, although it is difficult to demonstrate synergy at earlier time points in Figure 9 due to the ineffectiveness of HBV alone, lower doses of HBV did not show synergy with kanamycin at these earlier time points. A synergistic effect is, however, seen at the 24 hour time point. Two explanations for this type of delayed effect are proposed: (1) elimination of resistant mutants or (2) extension of the half-life of the antibiotic.

Decreasing probability of selection of resistant strains.

If both the honey bee venom and the antibiotic are present in a bacterial culture at bacteristatic doses, the probability that a resistant bacterium will survive the combined treatment is equal to the product of the probabilities that one would exist and survive both treatments. This would appear as a shifted synergistic effect, just as many generations of mutants would be required to multiply to a level detectable by OD readings. Mutant selection would appear as sporadic occurrence of drastically higher OD readings among replicate samples, which would be reflected by the standard deviation of the treatment. For example, when the effects of HBV treatment alone on kanamycin-resistant S. aureus were evaluated with kanamycin, the mean OD of the 12-hour point on the venom control curve was 0.65 with a standard deviation of 0.51 (Table A-8), indicating widely different readings at this time point. Thus, it may be possible that the synergistic effect observed here at the 24-hour point is the result of suppression of the HBV venom-resistant mutants.

Increasing antibiotic stability.

A possible mechanism of protection of the antibiotic from degradation cannot be ruled out. A common technique for increasing the effectiveness of antibiotics is to structurally alter the antibiotic to make it more stable in solution or more resistant to enzymatic attack. These types of alterations have been incorporated into many derivatives of the penicillin family of antibiotics. For example, penicillin V has a phenoxymethyl substitution, which creates a steric hindrance that protects the β-lactam ring of the antibiotic from enzymatic attack (Volk, 1978c). Such substitutions may also protect this end of the molecule from cyclization with the lactam ring, making the molecule more resistant to acid hydrolysis. These types of changes could cause a synergistic effect that is only detectable at bacteriostatic doses, since the antibiotic could not be more effective initially and the extended lifespan of the antibiotic would only be detectable if the bacterial culture did not reach a food-limited OD at that time. However, if HBV could cause such a change, more consistent results would be expected in repeat samples.

Evaluation of the MIC test

The checkerboard titration assay, which was developed to test HBV/antibiotic synergy, was time-consuming to apply in a broad study of the effects of HBV on different antibiotics and different bacteria. However, such a study was necessary to determine the trends among antibiotic classes against synergy with HBV, as well as to determine the spectrum of susceptibility among bacterial species to specific synergistic combinations of antibiotics and HBV. The modification of the automated MIC assays was made to facilitate this type of study.

Because of the limitations of the automated MIC test, the interpretation of the results was somewhat empirical. The results cannot be demonstrated as synergistic as opposed to additive interactions, since the effect of HBV alone was only recorded as inhibitory or non-inhibitory (weakly inhibitory doses of HBV would have been recorded as non-inhibitory, although an increase in MIC could actually be the result of an additive effect). However, in most tests, only certain antibiotics showed a decreasing MIC, indicating that the HBV dose was not additive. Thus, supporting the results of the checkerboard titration system, the use of these MIC tests should reliably indicate which antibiotic/HBV combinations have the greatest potential for specific groups of bacteria. In this regard, the MIC is used to direct future research.

Identification of the active honey bee venom component

Although the results of these studies suggest that the synergistic activities of honey bee venom are entirely contained in the melittin fraction, these results should be interpreted with caution. It is possible that small peptides or non-staining (Coomassie blue) compounds co-migrate with melittin in the chromatography due to ionic and hydrophobic interactions with melittin molecules. Melittin migrates as an aggregate that has five times the normal molecular weight in both native polyacrylamide gel electrophoresis and Sephade gel chromatography (Haberman, 1972). These small micelles could carry smaller hydrophobic compounds through the chromatography. Analyses to detect such types of contamination in the melittin fraction are also included and are discussed in Chapter 6.

As noted above, additional tests were conducted to demonstrate that HBV is also effective in increasing the activity of the fourth group of antibiotics referred to above, which is represented by rifampicin. The data are presented in Tables 6, 7 and Figures 16, 17.

The activity of Hymenoptera venom and other than HBV was also determined for bumblebee venom, wasp venom and bald faced hornet venom as shown in Tables 8, 9 and in Figures 18 and 20.

Some of the analogues mentioned above were also tested to determine their relative activities with respect to native melittin. The relative Activity is calculated as follows:

Melittin dose/analogue dose required to demonstrate equivalent synergy with polymyxin B x 100

The results obtained are shown in Table 10.

It appears that analogues in which the NH2-terminal end consists predominantly of basic amino acids are more active than analogues which have an NH2-terminal end and consist predominantly of neutral and/or acidic amino acids.

In vivo experiment introduction

In vivo experiments show that melittin, the main peptide of honey bee venom, increases the efficacy of a tested antibiotic, polymyxin B. Bacterial sepsis was developed in mice as a disease model. For the experiments, activities of polymycin B and melittin are compared separately and in combination against E. coli septicemia in two basic sets of experiments. In both sets of experiments, a synergistic interaction between melittin and polymyxin B is evident and is statistically confirmed by contrasting the treatment agents in each difference analysis. Thus, the ability of melittin to increase the efficacy of polymyxin B and to obtain antibacterial activity in vivo. Numerous citations to the literature cited above in the section entitled Background and Prior Art involve the use of honey bee venom, and in particular melittin, as an antimicrobial agent. However, these citations demonstrate only the in vitro efficacy of honey bee venom or melittin.

Several other systems have also made use of melittin as an artificial agent to disrupt various immune responses in in vitro isolate systems. Goodman et al. (1984) described B cell activation by melittin in vitro. Two other reports, one by Kondo and Kanai (1986) and one by Kondo (1986), describe the in vitro use of melittin to stimulate the bactericidal activity of membranes isolated from phagocytes from both mice and guinea pigs. Finally, one publication (Somerfield et al. 1986) relating to the effect of honey bee venom on the immune system describes the inhibition of neutrophil O production by melittin. Somerfield et al. suggest a role for melittin as an anti-inflammatory agent. This activity would most likely stimulate antibacterial defense in vivo.

Despite the substantial amount of research with melittin, it has not yet been demonstrated that melittin is active in vivo against infectious organisms. Importantly, no one has suggested the need for, or benefit of, the interaction of melittin. The results reported here demonstrate the beneficial interactions between melittin and polymyxin B when used in vivo to treat mice suffering from suffering from bacterial blood poisoning caused by E. coli.

Material and method

Female Swiss CD-1 mice (Charles River) were received weighing 18 to 20 grams. The mice were maintained at 77 +/-1 degrees Celsius and 30-45% relative humidity with a daily photoperiod of 12 hours. Upon arrival, each shipment of mice was maintained during a two-week acclimation period prior to their use in the experiment.

Polymyxin B (Sigma Chemical Company) was purchased in powder form with an activity of 7900 units/mg. A stock solution of 0.1 mg/ml in 0.85% NaCl was prepared and frozen at -20ºC in 5 ml aliquots until use.

Honey bee venom (HBV) was provided by Vespa Laboratories, Inc., Spring Mills, PA. HBV was the source of melittin, which was isolated using gel filtration as described above for the in vitro experiments. Melittin was quantitatively recovered by the Lowry protein assay (Lowry et al. 1951) and then freeze-dried. The freeze-dried melittin was reconstituted in 0.85% NaCl to a concentration of 0.1 mg/ml and frozen at -20ºC in 1.5 ml aliquots until further use.

E. coli strain #G1108E was obtained from the E. coli Reference Center of the Pennsylvania State University, University Park, PA. An overnight 5.0 ml Trypticase soy broth culture was used to inoculate 800 ml of fresh Trypticase soy broth. The Culture was grown overnight with gentle shaking. 200 ml of sterile glycerol was added to the culture and was then aseptically dispensed into 5.0 ml aliquots with stirring. These aliquots were frozen and stored at -20ºC. After thawing, each aliquot contained (+/-3) x 10⁸ viable bacteria/ml. The culture was then diluted 1:400 with Trypticase soy broth containing 2.5% gastric mucin (Sigma Chemical Company) prior to inoculation.

Mice were infected by intraperitoneal injection of 0.25 ml of a 1:400 dilution of bacteria (approximately 500,000 viable bacteria) suspended in trypticase soy broth with 2.5% mucin.

Before injection, polymyxin B and melittin were thawed, filter sterilized and conveniently diluted with 0.85% NaCl in such a way that the required dose was contained in 0.2 ml of solution. Thirty minutes after infection, this volume was then administered to mice by subcutaneous injection into the skin fold at the base of the neck. The skin fold is formed between the thumb and index finger in a basic movement-restricting grip.

Blood bacterial levels were determined from blood samples obtained by aseptic cardiac puncture. After cardiac puncture, the needle was removed from the heparin-coated syringe and 0.2 ml of blood was transferred to a tube containing 0.2 ml of 0.85% NaCl and mixed well. All samples were kept on ice until used. Two distribution plates for each sample at appropriate dilutions were prepared with trypticase oja agar and incubated at 37ºC overnight. All plates containing less than 400 colonies were were counted and recorded.

Results

In the first experimental design, four random groups of four mice were inoculated with E. coli as described above. Thirty minutes later, the four mice in each group were each treated with 0.2 ml of 0.85% NaCl solution containing one of the following: 1) 0.85% NaCl alone ("no treatment"); 2) 2.0 ug polymyxin B; 3) 50 ng melittin; or 4) 2.0 ug polymyxin B + 50 ng melittin. Twenty-one hours after the initial inoculation, blood samples were taken and the bacterial count per ml of blood was calculated by averaging the results of two-fold plate counts of the corresponding blood dilution (Table A-11). This experiment was performed in triplicate and the mean bacterial count per ml of blood was compared for each of the four treatments (Figure 21). A p-value of 0.0015 for the treatment effect in a two-way variance (adjusted for unequal sample sizes) indicated a significant difference in at least one of the treatment agents. Tukey's multiple comparison of means showed that the only agent that was significantly different was the agent of the group receiving the combination of 2 µg polymyxin B and 50 ng melittin. Comparison of the sum of the activities obtained by melittin and polymyxin used individually with their activity when used in combination confirmed by contrast in the analysis of differences that the interaction was indeed synergistic (p-value = 0.0493).

A second experimental design tested the effect of repeated treatments. Again, Four groups containing four mice were inoculated with E. coli and 30 minutes later given the same four treatments: 1) 0.85% NaCl alone ("no treatment"); 2) 2.0 ug polymyxin B; 3) 50 ng melittin; or 4) 2.0 ug polymyxin B + 50 ng melittin. Eighteen hours after the initial infection, each mouse was again challenged with the same E. coli inoculum and 30 minutes later given the same antibiotic/mellitin treatment. Five hours later (23 hours after the initial infection), blood samples were taken from each mouse to determine the number of bacteria in the blood. This experiment was repeated five times and the results (Table A-12) were evaluated by analysis of variance. These analyses showed that a significant difference (p-value = 0.001) was present in at least one treatment. Tukey's multiple comparison of means showed that repeated polymyxin B treatments resulted in a significant decrease in the number of bacteria per milliliter of blood. More importantly, Tukey's comparison showed that the number of bacteria in the blood of animals treated with polymyxin B plus melittin was significantly lower than the number in the blood of animals treated with polymyxin B or melittin alone (see Figure 22). A contrast within the analysis of variance indicates a high degree of confidence for the synergy of these two compounds (p-value = 0.0007).

Conclusions

The above experiments clearly demonstrate a synergistic interaction between the antibiotic polymyxin B and melittin. It is highly likely that melittin has the therapeutic Effects of other pharmaceutical agents due to its antimicrobial properties and its property of increasing membrane permeability.

The results of the first set of experiments (Figure 21) lacked statistical evidence for the effect of melittin alone, but comparisons of the absolute means lead to the conclusion of positive effects with melittin treatment alone. The results of the second set of experiments (Figure 22) again lacked a significant difference for treatment with melittin alone. Comparison of the absolute means of treatment leads to the conclusion of a negative effect of melittin alone. Recalling that the mice in the second set of experiments received double doses of melittin, this leads to the conclusion that the higher doses of melittin, when used without antibiotics, could exacerbate the infectious process. Previous experiments with higher doses of melittin verify this assumption. It is likely that the deleterious activity is present when melittin is used alone. Importantly, the effective use of melittin to treat infections was not found in a literature search. Antibiotics obviously counteract the negative effects of melittin, thus making combined therapy a significant development.

Antibiotic synergy demonstrated with a synthetic melittin analogue

The synergy observed between antibiotics and melittin can be achieved by replacing melittin with synthetic peptide analogues. Such analogues have been developed and synthesized for this purpose. In parallel tests with natural melittin, an equivalent antibiotic improvement is achieved.

Introduction

The analogue 6, with the structure shown below

was previously tested in vitro for synergy with polymyxin B against E. coli. This peptide varies from melittin at amino acids 22 and 24 (underlined) where the arginines were replaced with lysines. When used in the aforementioned tests, it showed an activity equivalent to that of natural melittin.

Materials and methods

Melittin was isolated from whole honey bee venom (Vespa Laboratories, Inc.) by gel chromatography, quantified by the Lowry protein assay, and stored freeze-dried. For these assays, lyophilized melittin was reconstituted to 0.4 mg/ml with distilled water, filter sterilized, and stored in 4.0 ml aliquots at -20ºC until use.

Analogue No. 6 was synthesized by Dr. Torben Saermark (Protein Laboratory, University of Copenhagen, Sigurdsgade 34, DK-2200 Copenhagen N, Denmark) and was determined to be better than 98% pure based on the elution profile of high pressure liquid chromatography on a C18 column using a 0-80% acetonitrile gradient in 0.1% trifluoroacetate was used. The peptide was obtained in lyophilized form and was reconstituted to approximately 0.2 mg/ml in 0.85% NaCl, filter sterilized and stored in 0.5 ml aliquots at -20ºC before use.

Polymyxin B (Sigma Chemical Company) with a specific activity of 7900 units/mg was reconstituted to 240 units/ml distilled water, filter sterilized and stored in 4.0 ml aliquots at -20ºC until use.

E. coli strain #G1108E was obtained from the Pennsylvania State University E. coli Reference Center (105 Henning Building, University Park, Pennsylvania, 16802). Inoculum was prepared from a culture grown in trypticase soy broth to mid-log phase. Sterile glycerol was added to make a final concentration of 20% and the culture was aliquoted and frozen at -20ºC until use.

A checkerboard titration synergy assay was performed, testing natural melittin and analog #6 in parallel with polymyxin B against E. coli. Equivalent doses of natural melittin and analog #6 were based on a Lowry protein assay performed simultaneously on aliquots of each after the final filtration. Both peptides were tested in the synergy assay at final media concentrations of 5 ug/ml and 10 ug/ml. Polymyxin B was tested at final media concentrations of 3 units/ml and 6 units/ml against both concentrations of both peptides.

Results

Synergy was best demonstrated with both the natural melittin and the #6 analog when tested at 10 ug/ml against #6 units/ml polymyxin (Table 11). Under these conditions (see Figure 23), the data from each peptide were statistically analyzed for synergistic activities at each time point. Using statistical contrasts, the mean activities of each peptide alone and polymyxin B alone were compared to the activity of the corresponding peptide-polymyxin B combination. Synergy was demonstrated at the 4, 6, and 8 hour time points for both melittin and the #6 analog (p-values = 0.0001).

In addition, the synergy curves for melittin (10 ug melittin + 6 units polymyxin B) and analog #6 (10 ug analog #6 + 6 units polymyxin B) were compared at each time point at different activity levels. At no time point could a significant difference be observed between the two curves.

Conclusions

The results show that Analogue #6, a synthetic melittin analogue, has very similar activity to melittin in terms of its capacity to enhance polymyxin B. Although the 12 hour time point leads to the conclusion that Analogue #6 has slightly better activity with polymyxin B than melittin, this difference in activity is minimal considering the differences in the amounts used in the peptides tested. By comparing the synergy difference produced by 10 ug melittin versus 10 ug analogue #6 to the synergy difference produced by 10 ug melittin versus 5 ug melittin (Table 11), it was found that the difference between the specific activities of melittin against analogue #6 was less than 10%. This research demonstrated that it is possible to synthesize melittin analogues with synergistic capabilities equal to or better than melittin.

Synergistic antibacterial activity of melittin and polymyxin B: Relative activities of melittin analogues

The synergy observed between antibiotics and melittin can also be achieved by replacing melittin with either synthetic analogues or chemically modified derivatives of the natural peptide. Synthetic melittin, five synthetic peptide analogues and a chemical modification of natural melittin were tested for growth inhibition of E. coli by synergistic interaction with polymyxin B. Their relative activities were compared with those of melittin from natural honey bee venom. Each peptide demonstrated synergistic interaction with polymyxin B. However, the specific activities differed significantly. Several analogues demonstrated synergistic activity superior to that of natural melittin.

introduction

Two types of melittin analogues, synthetic peptides and a chemical modification of natural melittin, were tested in vitro for synergy with polymyxin B in an antibacterial activity assay. Synthetic analogs include a group of synthetic peptides, each of which differs by two or more residues from the 26 amino acids of the melittin sequence. The chemical modification of melittin, NPS-melittin, consists of adding a -nitrophenylsulfenyl group to the tryptophan residue number 19 of natural melittin. The activity of each of these analogs was compared to the activities of natural and synthetic melittin. While each of these analogs showed some synthetic interaction with polymyxin B in vitro, significant differences in peptide activities were evident. These differences define the key properties of the melittin molecule in its role as an enhancer of polymyxin B activity.

Materials and methods

Natural melittin was isolated from whole honey bee venom (Vespa Laboratories, Inc.) by gel filtration chromatography, quantified by the Lowry protein assay, and stored freeze-dried. For these assays, freeze-dried melittin was reconstituted to 0.34 mg/mL with distilled water, filter sterilized, and stored in 4.0 mL aliquots at -20ºC until use.

NPS-melittin was synthesized from natural melittin by reacting the peptide with -nitrophenylsulfenyl chloride (NPS-Cl) as described for adrenocorticotropin by Ramachandran et al. The peptide was precipitated from a solution with ethyl acetate, resuspended in 0.1 N acetic acid and then passed through a Sephadex G-10 (LKB-Pharmacia, Piscataway NJ) column to remove the remaining NPS-Cl salts. Determination of the molecular absorbance of the derivative at 365 nm showed that more than 95% of the melittin was modified.

Synthetic melittin was purchased from Peninsula Laboratories, Belmont, Ca. A 0.5 mg sample was reconstituted to 0.3 mg/ml in 0.85% saline based on a Lowry protein assay. This sample was stored at -20ºC until use.

Synthetic analogues were synthesized by Dr. Torben Seamark (Protein Laboratory University of Copenhagen, Sigurdsgade 34, DK-2200 Copenhagen N, Denmark). Each analogue was tested for purity and was found to be better than 98% pure based on the elution profile of chromatography from a C-18 column using a 0-100% acetonitrile gradient. Each peptide was obtained in lyophilized form and was reconstituted to approximately 1.0 mg/ml in 0.85% NaCl, filter sterilized and stored in 1.8 ml aliquots at -20ºC until use. Each peptide solution was quantified by the Lowry protein assay prior to use. The Lowry results agreed well with the concentration estimates based on the dry weight of the peptide.

Polymyxin B (Sigma Chemical Company, St. Louis, Mo.) with a specific activity of 7,900 units/mg was reconstituted to 240 units/mg in 0.85% saline, filter sterilized, and stored in 4.0 ml aliquots at -20ºC until use.

E. coli G1108E was obtained from the Pennsylvania State University E. coli Reference Center (105 Henning Building, University Park, Pa. 16802). Inocula were obtained from a culture grown to mid-log phase in trypticase soybean broth. Sterile glycerol was added to the culture added to a final concentration of 20% and dispensed into 5.0 ml aliquots and stored at -20ºC until use. A checkerboard titration synergy assay was performed to test each peptide with polymyxin B against E. coli in parallel with melittin. Equivalent doses of natural melittin and each analogue were based on the Lowry protein assay, which is performed on aliquots of each peptide after a final filtration of the stock solution. All peptides were tested by the synergy assay with final concentrations in the medium of 5 ug/ml. The concentration of polymyxin B in the medium of all assays was 6 units/ml

Results

Table 12 shows the amino acid sequences of synthetic melittin analogues. For each synthetic analogue, the first twenty N-terminal amino acids were the same as those of natural melittin. Changes occurred in the six C-terminal amino acids and were indicated in bold.

Table 13 contains the growth curve readings for each of the compounds tested for synergistic interaction with polymyxin B. The value for each time point represents the mean and standard error of the mean of six samples. The "control" curve represents the growth of the culture without polymyxin B or peptide added. The effect of each peptide alone on the culture is not included in the table. However, these peptides, like melittin, have no effect on the growth of E. coli when used alone at 10 ug/ml or less. Thus, the "control" is also a representation of the culture when used with each peptide was treated alone.

When the growth curve of bacterial cultures treated with polymyxin B (6 units/ml) alone is compared to the curve of cultures treated with polymyxin B plus melittin (5 ug/ml), increased antibacterial activity is demonstrated as is an increase in the time required for the culture to overcome the treatment and reach log phase growth (Figure 24). Since treatment of the culture with melittin alone at 5 ug/ml would produce a growth curve that would essentially overlay the "control" curve, an increase in the time required for the culture to escape polymyxin B inhibition and reach mid-log phase in the presence of the peptide is evidence of the synergistic activity of the peptide. Thus, a shift of the growth curve presenting polymyxin B with peptide to the right of the curve representing treatment with polymyxide B alone is evidence of synergy.

The ratio of polymyxides B to melittin to bacteria in these experiments was set to produce minimal synergy so that the increased activity of the peptide analogues would be evident. Such increases are evident in Figure 24 for both synthetic melittin and NPS-melittin. All peptides were at equal concentrations and were determined by the Lowry assay.

Similar growth curves determining the synergistic activities of synthetic melittin analogues with polymyxin B are shown in Figure 25.

To illustrate the differences in melittin/analogue synergy with polymyxin B more clearly, Figures 24 and 25 were used to show the additional to calculate the time elapsed until each growth curve reached mid-log phase due to the addition of melittin or analogues, compared to treatment with polymyxin B alone. These values are presented in a bar graph in Figure 26. A Tukey range test was performed on the data in Table 13 to compare the readings obtained at each time point between the different treatments. The peptides were then grouped depending on their ability to show significantly different levels of growth inhibition (α = 0.05) for at least the time points of a growth curve. These groupings were designated by different markers in Figure 26.

Conclusions

These results demonstrate that a number of melittin analogues can be obtained by both amino acid substitutions and chemical modifications. These types of modifications can either increase or decrease the peptide-related synergistic activity. Based on Figure 26, the relative in vitro activities of melittin and its analogues can be given in ascending order as follows:

1 Analogue #7 A

2 natural melittin B

3 NPS-Melittin C

4 Analogue #6 C

5 synthetic melittin C

6 Analogue #2 C

7 Analogue #4 D

8 Analogue #4 D

Peptides with significantly different (α = 0.05) synergistic activities at the 5 ug/ml level are designated by letter groups.

It should be noted that the parameter used to establish potency, which shifts the time to mid-log phase of culture, increases stoichiometrically with the amount of peptide only over a narrow range of melittin increase. Since the limits of the linear range may not be equivalent for each analogue, the data presented here can only be used to determine the relative order of potency of these peptides at given concentrations and cannot be used to estimate quantitative differences. However, this order of potency does not allow the conclusion that the synergistic activity of the peptides is due to the number and exposure of positively charged side chains on amino acids in the C-terminal region.

Although the relative potency of these melittin analogues has been established in vitro, significant differences may exist in vivo. In vivo parameters such as absorption and clearance from the host may significantly alter this order of action in practical use. Side effects must also be considered. While the adrenocorticotropic activity of melittin is well documented, NPS melittin may have less adrenal activity than natural melittin because the NPS derivative of adrenocorticotropin (ACTH) induces 100 times less lipolytic activity than unmodified ACTH. For these reasons, any of the analogues included in this study should be considered for in vivo evaluation. Table 1 Concentrations of antibiotic and bee venom stock solutions tested against three different bacteria Organism Venom Ampicillin Kanamycin Polymixin E. coli S, aureus S. aureuskana Table 2 Representation and distribution of bee venom and antibiotics in a titration, checkerboard titration test Antibiotic dilutions Control BEE VEMONT DILUTIONS a = numerator, dilution level of the HBV stock solution b = denominator, dilution level of the antibiotic stock solution c = test position in a sequential arrangement of 75 reagent tubes Table 3 Contents and distribution of each component of the checkerboard titration test Tube No. Antibiotic Poison Bacteria Table 4 Effect of 4 ug/ml HBV on the MIC of eleven antibiotics against 8 gram-positive organisms Penicillin Methicillin Ampicillin Cephalothin Gentamicin Kanamycin Erythromycin Chloramphenicol Clindamycin Tetracycline Vancomycin 1 - Group "A" = 2 S. aureus strains 2 - Group "B" = 5 S. epidermidis strains 3 - Group "C" = one S. faecalis strain 4 - QC = S. aureus strain used for routine quality control testing of this test system 5 - A (-) indicates a MIC decrease of less than 2 dilutions 6 - A (+) indicates a MIC decrease of two or more dilutions Table 5 Effect of 4 ug/ml HBV on the MIC of 11 antibiotics on 4 E. coli strains E. coli strain Ampicillin Carbenicillin Piperacillin Cephalothin Cefoxitin Cefamandole Moxalactam Amikacin Gentimicin Chloramphenicol Tobramycin 1 - QC is an E. coli strain used for routine quality control testing of this test system. 2 - A (+) indicates a MIC decrease of 2 or more dilutions 3 - A (-) indicates a MIC decrease of less than 2 dilutions Table 6 Staphyloccus aureus Rifampicin=.01ug/ml or .001ug/ml Honeybee venom =4ug/ml Hours after inoculation Control Mean Rifampicin ug/ml Mean Venom ug/ml Mean Table 7 Pseudomonas aeruginosa Rifampicin=10ug/ml or 20ug/ml Honeybee venom =40ug/ml Hours after inoculation Control Mean Venom ug/ml Mean Rifampicin ug/ml Mean Table 8 Escherichia coli Polymyxin B=6.25Units/ml and 3.125Units/ml Bumblebee venom =5ug/ml and 20ug/ml (Megabombus pennsylvanicus) Hours after inoculation Control Mean BB Venom ug/ml Mean Table 9 Escherichia coli Polymyxin B =3.125 Vesp venom =5ug/ml (Vespula germanica) Hornet venom =5ug/ml (Dolichovespula maculata) Hours after inoculation Control Mean Table 10 Relative activity of analogues of protein or polypeptide components of Hymenoptera venom Analogue No. Relative activity Table 11 Mean optical densities (OD₆₆₀) of bacterial cultures versus time and treatment Control Mellittin - ug Analog ug Polymyxin B - Units Poly B - 6 U + Mel - 10 ug Table 12 Sequence of synthetic melittin analogues Natural melittin Analogue * Amino acids in bold represent changes from the native melittin sequence. Table 13 Optical densities (OD₆₆₀) of bacterial cultures for treatments versus time Hours after culture inoculation Control Polymyxin B Melittin + Pol B Synthetic + Pol B Analog + Pole B Table A-1 Results of the checkerboard test of ampicillin and bee venom against S. aureus. Time MeanA₆₆₀ Table A-1 (continued) Time MeanA₆₆₀ Table A-1 (continued) Time MeanA₆₆₀ Table A-2 Checkerboard test results of kanamycin and bee venom against S. aureus. Time Average A₆₆₀ Table A-2 (continued) Time MeanA₆₆₀ Table A-2 (continued) Time Mean A₆₆₀ Table A-3 Checkerboard test results of polymyxin B and bee venom against E. coli Time Mean A₆₆₀ Table A-3 (continued) Time MeanA₆₆₀ Table A-3 (continued) Time MeanA₆₆₀ Table A-4 Checkerboard test results of ampicillin and bee venom against E. coli Time AverageA₆₆₀ Table A-4 (continued) Time MeanA₆₆₀ Table A-4 (continued) Time MeanA₆₆₀ Table A-5 Checkerboard test results of kanamycin and bee venom against E. coli. Time MeanA₆₆₀ Table A-5 (continued) Time MeanA₆₆₀ Table A-5 (continued) Time MeanA₆₆₀ Table A-6 Checkerboard test results of polymyxin B and bee venom against E. coli. Time AverageA₆₆₀ Table A-6 (continued) Time MeanA₆₆₀ Table A-6 (continued) Time MeanA₆₆₀ Table A-7 Checkerboard test results of ampicillin and bee venom against kanamycin-resistant S. aureus. Time AverageA₆₆₀ Table A-7 (continued) Time MeanA₆₆₀ Table A-7 (continued) Time MeanA₆₆₀ Table A-8 Checkerboard test results of kanamycin and bee venom against kanamycin-resistant S. aureus. Time AverageA₆₆₀ Table A-8 (continued) Time MeanA₆₆₀ Table A-8 (continued) Time MeanA₆₆₀ Table A-9 Checkerboard test results of polymyxin B and bee venom against kanamycin-resistant S. aureus. Time MeanA₆₆₀ Table A-9 (continued) Time MeanA₆₆₀ Table A-9 (continued) Time MeanA₆₆₀ Table A-10 Results of equivalent doses of melittin and whole bee venom with and without kanamycin on S. aureus. Time Mean A₆₆₀ Table A-11 Rough data for the simple treatment model Log10 bacteria/ml blood Treatment Mouse elittin olymyxin B * No reading due to inappropriate blood sample. Table A-12 Rough data for the repeated treatment model Log10 bacteria/ml blood Treatment Mouse No treatment Mellitin Polymyxin

BIBLIOGRAPHY

Benton, A. W. 1965. Bee venom, its collection, toxicity and proteins. Thesis, Dept. Entomology, Cornell University, Ithaca, New York.

Benton, A. W., R. A. Morse and F. V. Kosikowski. 1983. Bioassay and standardization of venom of the honeybee. Nature 198:295-296.

Brangi, G. P. and M. Pavan. 1954. Bactericidal properties of bee venom (Translated title, in Italian). Isectes Social 1:209-217.

Brown, L. R., J. Lauterwein, and K. Wutlmich. 1980. High-resolution ¹H-NMR studies of self-aggregation of melittin in aqueous solution. Biochemistry Biophys. Acta 622: 231-244.

Carrizosa, J. and M. E. Levison. 1981. Minimal concentration of aminoglycoside that can synergize with penicillin in entrococcal endocarditis. Antimicrob. Agents Chemother. 20:405-409.

Coulson, C. C. and R. L. Kincaid. 1985. Gram-preparative purification of calmodulin and S-100 protein using melittin-sepharose chromatography. 69th Annual Meeting of the Federation of American Society for Experimental Biology. Federation Proceedings 44:1777.

Cynamon, M. H. and G. S. Palmer. 1983. In vitro activity of amoxicillin in combination with clavulanic acid against Mycobacterium tuberculosis. Antimicrob. Agents Chemother. 24:429-431.

Fennel, J. F., W. H. Shipman and L. J. Cole. 1968. Anti bacterial action of melittin, a polypeptide from bee venom. Proc.Soc. Exp. Biol. Med. 127:707-710.

Franklin, T. J. and G. A. Snow. 1981a. Biochemistry of antimicrobial action. Chapman and Hall, New York, New York. pp. 67-72.

Franklin, T. J. and G. A. Snow. 1981b. Biochemistry of antimicrobial action. Chapman and Hall, New York, New York. pp. 73-74.

Guralnick, M. W., L. M. Mulfinger and A. W. Benton. 1986. Collection and standardization of hymenoptera venoms. Folia Allergol. Immunol. Clin. 33:9-18.

Haberman, E. 1972. Bee and wasp venoms. The biochemistry and pharmacology of their peptides and enzymes are reviewed. Sci 177:314-322.

Haberman, E. and J. Jentsch. 1967. Sequence analysis of melittin from the tryptic and peptic cleavage fragments. Hoppe-Seyler's Z. Physiol. Chem. 340:37-39.

Hanke, W., C. Methfessel, H. U. Wilmsen, E. Katz, G. Jung. and G. Roheim. 1983. Melittin and a chemically modified trichotoxin form alamethicin-type multi-state pores. Biochem. Biophys. Acta 727:100-114.

Lauterwein, J., C. Bosch, L. R. Brown and K. Wuthrich. 1979. Physiochemical studies of the protein-lipid interactions in melittin-containing micelles. Biochemistry. Biophys. Acta 556:244-264.

Lauterwein, J., L. R. Brown and K. Wuthrich. 1980. High-resolution ¹H-MNR studies of monomeris melittin in aqueous solution. Biochemistry Biophys. Acta 622:219-230.

Lowry, O. H., N. J. Rosenbrough, A. L. Farr and R. J. Randall. 1951. Protein measurement with the folin phenol reagent. J. Biol. Chem. 193:265-275.

Moellering, R. C., C. Wennersten and A. N. Weinberg. 1971. Studies of antibiotic synergism against enterococci- J. Lab. Clin. 77:821-827.

Mollay, C. and G. Kreil. 1973. Fluorometric measurements on the interaction of melittin with lecithin. Biochemistry. Biophys. Acta 316:196-203.

Mulfinger, L. M., A. W. Benton, M. W, Guralnick and R. A. Wilson. 1986. A qualitative and quantitative analysis of proteins found in vespid venoms. J. Allergy Clin. Immunol. 77:681-686.

Ortel, S. and F. Markwardt. 1955. Investigations on the bactericidal properties of bee venom (Translated title, in German). Pharmacy 10:743-746. Abstracted in Chemical Abstracts. 1956. 50:1229c.

Schmidt-Lange, W. 1941. The bactericidal action of bee venom (Translated title, in German). Munchener Medizinische Wochenschrift 88:935-936.

Sebek, O. K. 1980. Antibiotics: volume 1: mechanism of action. D. Gottlieb and P. D. Shaw (eds.). Springer-Verlag, New York. pp. 142-149.

Tu, AT 1977a. Venoms: chemistry and molecular biology. John Wiley and Sons, Inc., New York, London, Sydney, and Toronto. pp. 1-16.

Tu, A. T. 1977b. Venoms: chemistry and molecular biology. John Wiley and Sons, Inc., New York, London, Sydney, and Toronto. pp. 501-512.

Tu, A. T. 1977c. Venoms: chemistry and molecular biology. John Wiley and Sons, Inc., New York, London, Sydney, and Toronto. pp. 505-509.

Volk, W.A. 1978a. Microbiology and physiology. C. May and J. Frazier (eds.). J. P. Lippincott Company, Phila., New York, San Jose and Toronto. pp. 121-122.

Volk, W.A. 1978b. Microbiology and physiology. C. May and J. Frazier (eds.). J. P. Lippincott Company, Phila., New York, San Jose and Toronto. pp. 122-126.

Volk, W.A. 1978c. Microbiology and physiology. C. May and J. Frazier (eds.). J. P. Lippincott Company, Phila., New York, San Jose and Toronto. pp. 130-133.

Volk, W.A. 1978d. Microbiology and physiology. C. May and J. Frazier (eds.). J. P. Lippincott Company, Phila., New York, San Jose and Toronto. pp. 133-135.

Yunes, R. A. 1982. A circular dichroism study of the structure of Apis melifera melittin. Arch. Biochem. Biophys. 216(2):559-565.

ADDITIONAL BIBLIOGRAPHY

Goodman, M. G. and W. Oh, Weigle. Regulation of B-lymphocyte proliferative responses by arachidonate metabolites: Effects on membrane-directed versus intracellular activators. J. Allergy Clin. Immunol. 74:418-425, 1984.

Kondo, E. and K. Kanai. Bactericidal activity of the membrane fraction isolated from the phagocytes of mice and its stimulation by melittin. Japane. J. Med. Sci. 39:9-20, 1986.

Kondo, E. Melittin-stimulated antimycobacterial activity of the membrane fraction isolated from the phagocytes of guinea pigs. Japane. J. Med. Sci. 39:21-24, 1986.

Somerfield, S. D., J. Stach, C. Mraz, F. gervais, and E. Skamene. Bee venom melittin blocks neutrophil Oz production. Inflammation 10:175-182, 1986.

Mulfinger, L. M. The synergistic activities of honey bee venom with antibiotics. Unpublished M. S. thesis, Pennsylvania State University, 1986 (contained in U. S. Patent Appln. S. N. 096,628).

Lowry, O. H., N. J. Rosenbrough, A. L. Farr, and R. J. Randall. Protein measurement with the folin phenol reagent. J. Biol. Chem. 193:265-275. 1951.

Ramachandran, J. and Virginia Lee. Preparation and properties of the -nitrophenyl sulphenyl derivative of ACTH: an inhibitor of the lipolytic action of the hormone. Biochem. Biophys. Res. Com. 38(3):507-512. 1970.

Scoffone, E., A. Fontana, and R. Rocchi. Sulphenyl halides as modifying reagents for polypeptides and proteins. I. Modification of tryptophan residues. Biochemistry 7(3):971-979. 1968.

Couch, T. and A. Benton. The effect of venom of the honey bee, Apis mellifera L., on the adrenocortical response of the adult male rat. Toxicon 10:55-62. 1972.

Claims (5)

1. Composition for treating infections in mammals, containing a first antibiotic agent effective against the infection and a second agent consisting of at least one hymenoptera venom, at least one active protein component of a hymenoptera venom, at least one polypeptide component of a hymenoptera venom, at least one similar or chemically modified derivative of an active protein component of a hymenoptera venom, at least one similar or chemically modified derivative of a polypeptide component of a hymenoptera venom, or mixtures thereof, wherein the second agent shows a positive biuret reaction and wherein the ratios between the first antibiotic agent and the second agent are selected such that the second agent enhances the activity of the first antibiotic agent. 2. Composition according to claim 1, in which the first antibiotic agent is an antibiotic from the group ampicillin, kanamycin, polymyxin B or rifampicin. 3. Composition according to claim 1 or 2, in which the second agent consists of honey bee venom, bumblebee venom, wasp venom, hornet venom, active protein components of at least one of these venoms, active polypeptide components of at least one of these venoms, similar or chemically modified derivatives of melittin, bombilitin, mastoporan and crabolin or mixtures thereof. 4. A composition according to claim 1, 2 or 3, wherein the first antibiotic agent contains ampicillin, kanamycin, polymyxin B or rifampicin and the second agent contains honey bee venom, melittin, similar or chemically modified derivatives of melittin, mastoporan, similar or chemically modified derivatives of mastoporan or mixtures thereof. 5. A composition according to any preceding claim, in which the first antibiotic agent contains ampicillin, kanamycin, polymyxin B or rifampicin and the second agent is honey bee venom or melittin. A dosage unit for treating an infection in a mammal containing an effective dosage of a medicament in the composition of any preceding claim.