In vitro elution and E. coli and T. pyogenes killing prole of penicillin, ampicillin, tetracycline, tulathromycin, and orfenicol from Plaster of Paris beads

Background: The objectives of this study were to create a protocol for making antibiotic impregnated Plaster of Paris (AI-PoP) beads using penicillin, ampicillin, tetracycline, tulathromycin, and orfenicol, to determine the in vitro elution rates of those antibiotics in the beads, and to quantify the in vitro inhibition of E. coli and T. pyogenes by the bead eluent in a proof-of-concept study using ow cytometry. The AI-PoP beads were made using Plaster of Paris powder, antibiotic, and tap water, cured for 24 hours, sterilized by ethylene oxide, and stored up to 5 months before testing. For each antibiotic, twenty beads were combined with bovine serum in sterile tubes and incubated at 37° C on a rocker. Serum was replaced at intervals over the 14 day study period, and antibiotic concentrations were determined by high pressure liquid chromatography with mass spectrometry. Separately, the growth of E. coli and T. pyogenes in eluent from 10 beads for each antibiotic was quantied by ow cytometry. Results: Antibiotic was detected in AI-PoP bead eluent for 14 days for all but the ampicillin beads, for which antibiotic was detected for 8 days. The concentration of antibiotic in eluent was greater than ve times the minimum inhibitory concentration (MIC) of tested bacteria for the entire study period for penicillin, tetracycline, and orfenicol. The concentration of ampicillin remained greater than ve times the MIC of E. coli for 4 days and T. pyogenes for 5 days. The concentration of tulathromycin remained greater than ve times the MIC of E. coli for 5 days and T. pyogenes for the entire study period. The CFU/ml of live E. coli and T. pyogenes was reduced over a 72-hour period by 1-3 log 10 CFU, with the exception of tetracycline, which reduced CFU/ml of T. pyogenes by less than log 10 CFU. Conclusions: AI-PoP beads containing penicillin, tetracycline,

Results: Antibiotic was detected in AI-PoP bead eluent for 14 days for all but the ampicillin beads, for which antibiotic was detected for 8 days. The concentration of antibiotic in eluent was greater than ve times the minimum inhibitory concentration (MIC) of tested bacteria for the entire study period for penicillin, tetracycline, and orfenicol. The concentration of ampicillin remained greater than ve times the MIC of E. coli for 4 days and T. pyogenes for 5 days. The concentration of tulathromycin remained greater than ve times the MIC of E. coli for 5 days and T. pyogenes for the entire study period. The CFU/ml of live E. coli and T. pyogenes was reduced over a 72-hour period by 1-3 log 10 CFU, with the exception of tetracycline, which reduced CFU/ml of T. pyogenes by less than log 10 CFU.
Conclusions: AI-PoP beads containing penicillin, tetracycline, tulathromycin and orfenicol elute antibiotic well above the MIC of tested bacteria for the two week study period and provide adequate local bacterial growth inhibition to reduce bacterial growth in vitro . These antibiotics show promise for delivery in joint and wound infections.

Background
Synovial sepsis, post-operative surgical site infections, and complicated wounds are a devastating cause of increased morbidity and mortality in large animal veterinary patients and can be career or even lifeending. [1][2][3] While systemic antibiotics are the mainstay of treatment, complementary local antimicrobial therapies are often employed. These adjunct therapies are especially useful in the treatment of osteomyelitis, synovial sepsis, and surgical site infections. [4][5][6] The bene ts of local therapy include the ability to achieve local drug concentrations many times above bacterial minimum inhibitory concentration (MIC), avoidance of high systemic doses of antibiotics, and cost effectiveness. [7][8][9] The options for local antimicrobial therapy include intravenous and intraosseous regional limb perfusion 10,11 and sustained release of antimicrobials eluted from implanted carriers such as collagen sponges, 12 polymethylmethacrylate, 13,14 plaster of Paris (PoP), 7-9 constant rate infusion systems, 15 and hydroxyapatite cement. 13 Regional perfusions can achieve antimicrobial concentrations many times higher than the MIC of common pathogens, however use is limited to the distal extremity due to the need to place a tourniquet proximal to the area of interest and because high concentrations of drug are generally not maintained for very long after tourniquet removal. Constant rate infusion systems are also able to achieve high antimicrobial concentrations, however breakage or blockage of the intrasynovial catheter is common. 15 Nonbiodegradable polymethylmethacrylate implants are easy to use, however removal may be necessary. In addition, creating the implant produces an exothermic reaction, thus only heat stable antibiotics may be used. 15 Biodegradable implants, including collagen sponges, hydroxyapatite cement, and PoP, offer the advantage of not requiring removal and being potentially more biocompatible. Puri ed type I collagen sponges are biocompatible, biodegradable, and are characterized by low immunogenicity. 12 Collagen sponges and hydroxyapatite cement have the disadvantage of being more expensive than PoP. Plaster of Paris, or calcium sulfate hemihydrate, is inexpensive, readily available, and easy to use for local antimicrobial delivery. 15 In addition, it is biocompatible, biodegradable, and osteoconductive when used in fracture repair and established osteomyelitis, providing a scaffold on which new bone formation can occur. 7,15 The in vitro elution of antibiotic from PoP beads has been reported for gentamicin, amikacin, clindamycin, enro oxacin, vancomycin, tobramycin, and cefazolin. [7][8][9]16,17 These studies show that the elution pro le is different among antibiotics.
The use of antibiotic impregnated PoP (AI-PoP) beads constitutes an extra-label drug use under the Animal Medicinal Drug Clari cation Act of 1994, which also prohibits or limits the use of previously studied drugs in food producing species. Antibiotics elected for use in this study include penicillin, ampicillin, tetracycline, tulathromycin, and orfenicol. These antibiotics were chosen because they are commonly used in food animal practice and extra-label use may be allowed.
The objectives of this study were to (1) create a protocol for making AI-PoP beads using penicillin, ampicillin, tetracycline, tulathromycin, and orfenicol, (2) determine the in vitro elution properties of the AI-PoP beads over 14 days, and (3) quantify the in vitro inhibition of one isolate each of E. coli and T. pyogenes by the AI-PoP bead eluent over 72 hours.
We hypothesized that (1) the elution of all antibiotics from the AI-PoP beads would have a rapid initial phase and complete elution within two weeks, (2) the initial concentrations of antibiotic in the eluent would exceed the MIC of bacterial isolates used, and (3) AI-PoP bead eluent would inhibit the growth of single isolates of E. coli and T. pyogenes bacteria as determined by ow cytometry.

Results
All AI-PoP beads eluted detectable concentrations of antibiotic for the 14-day sampling period, with the exception of ampicillin, which eluted detectable concentrations of antibiotic for 8 days. Mean observed peak ampicillin, orfenicol, and tetracycline concentrations occurred at four hours, mean observed peak tulathromycin concentration occurred at 12 hours, and mean observed peak penicillin concentration occurred at 24 hours (Figure 1, a-e).
The amount of antibiotic released in the rst 72 hours, expressed as a percent of antibiotic released over the two-week study period, was greatest for ampicillin and least for orfenicol ( Table 1). The total amount of antibiotic released over the two-week study period, expressed as a percent of total incorporated antibiotic, was greatest for orfenicol and least for tetracycline ( Table 1).
The MIC data for each antibiotic and bacterial strain, as provided by the diagnostic lab at our institution, is shown in Table 2. Some authors consider time dependent antibiotics, such as those used in this study, to be most e cacious when concentration is maintained 2-5 times higher than the minimum inhibitory concentration (MIC) for at least 50% of the dosing interval. 9 The concentration of antibiotic within the eluent reported at each time point was greater than ve times the MIC of both E. coli and T. pyogenes for penicillin, tetracycline, and orfenicol for the entire study period. The concentration of ampicillin remained greater than ve times the MIC of E. coli for 4 days and T. pyogenes for 5 days. The concentration of tulathromycin remained greater than ve times the MIC of E. coli for 5 days and T. pyogenes for the entire study period.
Ampicillin had the shortest elution half-life, and penicillin had the longest ( Table 3). The elution half-life of orfenicol and tulathromycin could not be calculated in any of the samples because the slope was not negative.
The results of the ow cytometry proof-of-concept study are shown in Figure 2 a-b. The decrease in concentrations observed at 24 and 48 hours of growth represents the decrease in concentrations that resulted from adding 1 ml of Mueller Hinton (MH) broth to the culture tube following sampling. The CFU/ml of E. coli is reduced by approximately 1-2 log 10 CFU by 72 hours for all antibiotics compared to the control sample. The CFU/ml of T. pyogenes is reduced by 3 log 10 CFU for all antibiotics except tetracycline by 72 hours compared to the control sample. CFU reduction in the tetracycline sample was less than log 10 CFU.

Discussion
This study was successful in creating antibiotic impregnated beads using ve different antibiotics. The rst hypothesis, that the elution of all antibiotics from the AI-PoP beads would have a rapid initial phase and complete elution within two weeks, was rejected. Only the elution of ampicillin was complete within two weeks; the remaining antibiotics eluted antibiotic for the full two-week study period. The second hypothesis, that the initial concentrations of antibiotic in the eluent would exceed the MIC of bacterial isolates used, was accepted. The concentration of antibiotic in the eluent exceeded ve times the MIC of bacterial isolates for a minimum of four days. The third hypothesis, that the AI-PoP bead eluent would inhibit the growth of single isolates of E. coli and T. pyogenes bacteria as determined by ow cytometry, was also rejected. The growth of T. pyogenes did not appear to be inhibited by tetracycline; however, the growth of both bacteria was inhibited by the remaining antibiotics.
Size and composition of AI-PoP beads have previously been shown to in uence elution pattern. 18 The amount of antibiotic eluted has been shown to be proportional to dissolution of the bead, with the rate of dissolution being affected by increases in porosity. 19 Porosity, determined by scanning electron microscopy, 19 and rate of bead dissolution were not analyzed in this study because the ideal porosity and other biomaterials properties for AI-PoP beads remains unknown. It has been speculated that factors such as storage temperature and humidity, length of time stored, and different antibiotics used may have an effect on bead dissolution.
This study is consistent with previous reports that found a variable volume of liquid is needed for individual batches in order to obtain an appropriate antibiotic-PoP mixture consistency prior to lling the bead mold. 7,9 Because of individual properties associated with each antibiotic, each AI-PoP bead preparation was different. Because procaine penicillin G was used, penicillin beads were cured and stored in a refrigerator in accordance with the handling instructions of the drug. It is unknown whether this was necessary, or if curing and storage at room temperature would have been acceptable.
A sampling period of two weeks was chosen in this study. Based on the rapid elution of antibiotic from AI-PoP beads observed in a previous study using PoP beads, 7 a similar elution pro le was anticipated.
However, we were surprised that in the current study, AI-PoP beads eluted antibiotic for the entire 14-day study period, with the exception of ampicillin. In previous studies using PoP beads, antibiotics such as gentamicin, amikacin, and clindamycin exhibited a rapid elution pro le, while vancomycin, enro oxacin, and tobramycin exhibited a prolonged release elution pro le. [7][8][9]16 The elution pattern of penicillin, ampicillin, and tulathromycin in our study is similar to the previously reported elution pattern of most antibiotics, 7,9 characterized by an initial rapid release of antibiotic, subsequently followed by low concentration release for the remainder of the study period. The elution pattern of tetracycline and orfenicol is atypical in that a relatively constant level of elution was maintained for the entire study period. This pattern has previously been reported for enro oxacin. 9 Previous studies have used eluent to qualitatively describe bacterial inhibition. 7,8 Flow cytometry has been shown to be an accurate and rapid method of determination of the concentration of bacteria in a sample while distinguishing between live and dead organisms. 20 This proof-of-concept portion of the study serves to show that ow cytometry can be used to quantitatively describe the effect of AI-PoP eluent on bacterial inhibition.
The choice of antibiotics for systemic administration should be based upon culture and antimicrobial susceptibility data, but because of a variety of factors including cost, time to achieve C&S results, negative growth on cultures, and other reasons, the choice is often based upon clinical experience, empirical data, and commonly known bacterial isolates associated with speci c conditions. The choice for local delivery of antibiotics is based upon similar criteria, although MIC breakpoints used to identify susceptible and resistant isolates are not predictive when antibiotic drugs are applied topically or locally, since achievable concentrations and effects on clinical outcome are unknown. Additional factors that could potentially in uence the performance of local antimicrobials include bacterial burden, wound contamination, and wound exudate. It remains unknown how this in vitro work is to relate in vivo. There are many factors of in vivo infection that were not replicated by this study, including microbe-related risk factors such as bio lm, and host-related factors such as an immune response.
The antibiotics chosen for this study were selected because they are labeled for use in food producing species and their extra-label use may be allowed. Food producing species are subject to restrictions on drug use such as provisions within the Animal Medicinal Drug Clari cation Act of 1994. It is recommended that the Food Animal Residue Avoidance Database (FARAD) be consulted for any extralabel drug use, including those described in this report. It is unknown what the serum or tissue concentrations of these drugs might be after application of AI-PoP bead, so caution should be used when predicting withdrawal times.
A limitation of this study is that elution was only investigated for two weeks. Elution data showed that a longer period would have been bene cial for penicillin, tetracycline, tulathromycin, and orfenicol. An additional limitation of this study is that using PoP powder and water from sources that are not regulated for purity may serve as potential sources of variability on both the application of the suggested recipes, as well as the elution of antibiotic from AI-PoP beads. Another limitation is that the materials properties of the beads were not analyzed, thus we were not able to provide further information regarding the rate of dissolution of the beads.

Conclusions
Future studies should focus on the in vivo application of ampicillin, orfenicol, penicillin, and tulathromycin AI-PoP beads to corroborate the results of this in vitro study. Ampicillin, orfenicol, penicillin, and tulathromycin appear to perform well in eluting antibiotic from AI-PoP beads and inhibiting the growth of E. coli and T. pyogenes. These antibiotics show promise for use AI-PoP beads in food animal species.

Methods
Based on methods previously described, 7,9 AI-PoP beads were made as follows: twenty grams of PoP powder was thoroughly mixed with antibiotic powder or liquid in a specimen cup. When needed, water was added to achieve a pourable mixture (see Table 4 for weights and volumes used). Ampicillin and tetracycline powder were used, whereas penicillin, tulathromycin, and orfenicol were in solution. All AI-PoP bead ingredients were chosen for being inexpensive and easily accessible to general practice veterinarians, as it was desired to create a bead recipe that could be replicated and used in a general practice setting. In this study, all PoP powder used was derived from the same carton, all water was obtained from the same source in the hospital, and all antibiotics were derived from the same bottle or vial. All bead types used experimentally were created in a single batch, thus eliminating the need to control for variability in lot numbers. Each mixture was poured into a silicone candy bead mold (Chicago Culinary FX, Westchester, IL) and a wooden tongue depressor was used to completely ll the wells. The use of a silicone bead mold allowed standardization of bead size in this study. The AI-PoP beads were cured for at least 24 hours at room temperature (with the exception of penicillin beads, which were cured at 4 °C), sterilized by ethylene oxide, and stored at room temperature (with the exception of penicillin beads, which were stored at 4 °C based on storage conditions identi ed on the drug label). Testing was completed within ve months of AI-PoP bead manufacturing.
For each antibiotic, twenty AI-PoP beads and seven milliliters of bovine serum (Lampire Biological Laboratories, Inc., Pipersville, PA) were aseptically transferred to individual sterile polypropylene test tubes. The number of beads, choice of serum, and volume of serum were chosen based on previous studies 7, 9 and because it was su cient to immerse the beads in serum. The tubes were maintained in a rotating incubator at 37 °C, except when removing eluent and adding serum. The entire volume of eluent was removed at speci ed time points and replaced with fresh bovine serum ( Table 5). As previously described, 7 the volume of serum was decreased to 5 mL at four days to re ect the decrease in effusion commonly observed in clinical patients. Eluent samples were stored at -70 °C until analyzed (approximately one month). The eluent from each AI-PoP bead type was tested in quadruplicate; four test tubes of AI-PoP beads were tested for each antibiotic.
Concentrations of ve antibiotics were determined using high-pressure liquid chromatography with mass spectrometry detection (LC-MS/MS) after protein precipitation of samples with acetonitrile. LC-MS/MS was performed using an Agilent 1100 Pump, column compartment, and autosampler (Santa Clara, CA, USA) coupled to an ion trap mass spectrometer (LTQ, Thermo Scienti c, San Jose, CA, USA). Serum samples, serum spikes, serum quality controls, and bovine serum blanks, 50 or 100 µL, were mixed with 450 or 400 μL of acetonitrile to precipitate serum proteins. The acetonitrile contained an internal standard at a concentration of either 250 or 1000 ng/mL. The samples were vortexed for 5 seconds and centrifuged for 10 minutes at 7500 rpm to sediment the protein pellet. A portion, 20 or 50 μL, of the supernatant was then diluted with water to 1 mL of volume in an autosampler vial. Further dilution of the samples, if needed, was done with blank serum that had been precipitated and diluted identically as the samples. These dilutions were done in autosampler vials tted with glass inserts and dilutions were 1:5 or 1:10. The vials were then centrifuged at 2,400 rpm prior to analysis. All the quality control samples passed the tolerance of being within ± 15% of the nominal concentration. The limit of quantitation of the analysis was 0.5 μg/mL with a limit of detection of 0.1 μg/mL.
The antimicrobial activity of eluent was evaluated against two bacterial isolates, an American Type Culture Collection (ATCC) strain of E. coli and a clinical isolate of T. pyogenes. These are two of the most common isolates in cattle from cases of synovial sepsis. 21 Together they encompass a broad spectrum of bacteria, T. pyogenes being a slow growing Gram-positive facultative anaerobe, and E. coli being a fastidious Gram-negative facultative anaerobe. 22,23 Using the McFarland standard, colonies of pure cultures of E. coli and T. pyogenes were used to inoculate test tubes containing ten AI-PoP beads and ve milliliters of MH broth to a concentration of 10 7 CFU/mL. This concentration was elected because it was the most concentrated inoculum that could be reasonably achieved using these methods. It has been shown that the risk for surgical site infection is markedly increased when contaminated with more than 10 5 microorganisms per gram of tissue. 24 Thus, the inoculation dose of 10 7 was considered su cient to represent clinical infection. A single negative control test of MH broth lacking AI-PoP beads was also inoculated at the same concentration for each bacterium. Following inoculation, the tubes were maintained in a rotating incubator at 37 °C. Twenty-four, 48, and 72 hours after inoculation, a one milliliter sample of eluent was obtained for quanti cation of live and dead bacteria by ow cytometry, and the volume was replaced with MH broth. In this proof-of-concept phase of the study, a single test of each AI-PoP and bacteria combination was completed. Live and dead bacteria in samples were stained for quanti cation by ow cytometry using a commercially available kit (Molecular Probes, Inc.; Eugene, OR). Samples were prepared according to manufacturer's instruction, then immediately transported to the ow cytometry laboratory for quanti cation. The number of events in the bead region was set to 5,000 and the sample was assayed at 488 nm. The data were processed by framing the region around the live and dead populations to yield events data for live and dead bacteria. The bacterial culture density was calculated using outcome events data and dilution factors. Could not be calculated because slope was not negative Tetracycline Could not be calculated because slope was not negative