As we described on the Product Design page, the risk of environmental escape limits the use of genetically modified organisms. Environmental escape not only is a risk that should be avoided out of public and environmental safety reasons, but also contributes to public fear of GMOs (1). We consulted several RIVM experts who confirmed that escape risk is a limiting factor in the advancement of GMO technologies. Environmental safety threats related to GMO escape includes the damaging of ecosystems, the risk of horizontal gene transfer and effects on biodiversity (2).

Gelcatraz has been designed with the intention of minimizing escape risk. We aim to provide a device which makes working with GMOs inside and outside of the laboratory more convenient and safer. Our adhesion system prevents bacterial leakers by securely anchoring E. coli to our device. Experimental results indicate that our adhesin decreases leakage by several orders of magnitude. Although leakage is low thanks to the adhesin, it is not yet zero. To minimize the consequences of a leakage event, we designed a kill switch for our E. coli.

To research the problem of antibiotic resistance and the need for antibiotics reduction, we approached several experts. All of them highlighted the urgency of the problem. Plasmacure as well as the doctors in the Maasstad Hospital explained that S. aureus and P. aeruginosa are increasingly becoming resistant against commonly used antibiotics. This makes treatment more difficult and even puts patients’ lives at risk. We also learned that the use of systemic, unspecific antibiotics contributes to the development of antibiotic resistance, as described under Product Design .

To contribute to a solution for the treatment of antibiotic resistant infections, we designed our living material to secrete bacteriocins as an alternative to antibiotics. Details on their functioning can be found here. Immunity to lysostaphin has not been commonly observed, but it possible (11). We aim to reduce this risk by integrating a quorum sensing system in the design. This way, the relevant bacteriocins will only be secreted in the presence of S. aureus and/or P. aeruginosa.

Cécile van der Vlugt, risk assessment expert at the RIVM, told us that the use of antibiotics in research and development is not an urgent problem, as the quantities used are low, and the types of antibiotics use are generally basic ones. For those antibiotics, many alternatives still exist in the unlikely case of resistance developed due to research work. Nevertheless, we asked professor Van Hest about the possibilities of replacing the commonly used antibiotic resistances with a dependence on artificial amino acids. He said that, albeit not yet used, dependence on artificial amino acids as a selection mechanism may be even more reliable than antibiotic resistance. The reason is that bacteria tend to get rid of genes that are temporarily unused, e.g. in the absence of antibiotics. As genes related to artificial amino acid dependence generally encode for essential proteins, they do not run this risk.

In conclusion, our project offers a solution to the treatment of infections caused by antibiotic resistant bacteria. We already reflected on ways to prevent resistance against our own bacteriocins. Finally, we explored how we as scientists may limit the use of antibiotics in R&D to prevent the development of additional resistance.

Staphylococcus aureus is a pathogenic species. It is one of the most important infections-causing bacteria in humans and it is easily transmitted through human contact and via the air (3,4). The pathogenicity of S. aureus requires us to take preventive measures before handling it. As we did not have access to the ML2 lab facilities required for use of S. aureus, we contacted our partner PAMM for help. They performed the experiments with S. aureus for us. As a recognized diagnostic laboratory, they work with S. aureus on a regular base and know which precautions to take.

Our hydrogel production requires toxic catalysts and yields a hazardous intermediate product. How do we ensure a safe synthesis and polymerization process and prevent the presence of toxic intermediates in the final product? Good Manufacturing Practices (GMP) are regulations which ensure the safety and quality of food and medicine. As part of GMP, it is thus important to protect the chemical safety of Gelcatraz and the safety of its production process. Furthermore, chemical waste should be minimized, and the costs should preferably be low. Below, we provide the results from our research into these topics. In short, our research revealed that all toxic components are removed by extensive dialysis of the methacrylated dextran. Furthermore, a leakage test showed that no leakage occurs from the gels over a period of two weeks. The use of a novel, radiation-based crosslinking method and alternative catalysts could reduce the toxicity of the synthesis and polymerization reactions. The gel is biodegradable.

Toxic chemicals and intermediates

Dextran itself is biocompatible and safe, but its synthesis and polymerization require several toxic chemicals. Furthermore, some toxic intermediates are produced. While handling these chemicals, we take all required precautions such as working in the fume hood and wearing protective gloves. However, it would be even better to replace them by less toxic equivalents. We researched replacement options for two steps: the synthesis of methacrylated dextran and polymerization of methacrylated dextran into gels.

Synthesis of methacrylated dextran

For the synthesis of the gel, we use some toxic chemicals, the most important being glycidyl methacrylate (GMA) and 4-dimethylaminopyridine (DMAP). GMA is carcinogenic but required to functionalize the hydroxyl groups of dextran. We handle it with syringes to avoid exposure. Furthermore, we remove all traces of GMA after the synthesis and confirmed this with H-NMR. The final product is thus completely safe. DMAP is a base that acts as a catalyst for the coupling of GMA to the dextran molecules. It is toxic if swallowed and fatal in contact with skin. In the absence of a catalyst, no GMA incorporation could be detected (5). However, we researched alternatives to catalyse the reaction. Triethylamine and pyridine are no reasonable alternatives due to the associated high conversion times (5). Another alternative for DMAP could be tri-ethylenediamine, also known as DABCO. DABCO is like DMAP an organic base and therefore it can theoretically catalyse the methacrylation reaction. However, unlike DMAP, DABCO possesses less toxic properties and DABCO is much safer to handle than DMAP. DABCO is from an economical point of view also a better choice as its commercially available for a lower price. A downside to using DABCO as a catalyst for this methacrylation reaction is that it is less basic than DMAP (6). As the deprotonation of hydroxyl groups by a base is the driving force for the reaction, the reaction time will most likely increase when DABCO is employed as a catalyst instead of DMAP. A comparison of the performance of DABCO in our reaction is required to make an informed choice between DMAP and DABCO.

Methacrylation of the non-hazardous dextran results in molecules with very toxic and carcinogenic properties due to the reactive methacryl groups (7). No material safety data sheet exist for methacrylated dextran. To safely handle this toxic viscous oil, it is kept in the fume hood at all times and only handled using piston pipettes or syringes. To purify the methacrylated dextran from the organic mixture, we dialyse it against demineralized water. To verify the complete removal of the DMSO, DMAP and GMA, the 1H nuclear magnetic resonance spectrum of the dialyzed reaction product was compared with the spectra of pure DMAP and GMA. After 72 hours of dialysis, trace amounts of unidentified by-products could be observed in the NMR spectrum. After one week of dialysis, no more trace amounts of any reactants or byproducts can be observed. The obtained spectrum matches with the one reported in the literature (8). The spectrum of purified dextran can be seen in the hydrogel results.

Polymerization of methacrylated dextran

To form the hydrogel, the methacrylated dextran polymers are crosslinked by radical polymerisation in water. The polymerisation is initiated by tetramethylethylenediamine (TEMED) and ammonium persulfate (APS). TEMED catalyses the homolytic scission of APS to generate radicals which crosslink the methacrylate groups, resulting in the formation of a hydrogel. APS and TEMED are both flammable and irritable. Because TEMED is a catalyst, we only work with small amounts of the substance (10 microliter at a time) using pipets, reducing the risks associated with this compound. TEMED is vital for the liberation of radicals from APS. In the absence of TEMED, polymerisation at room temperature by solely APS is slow, so at -20°C the polymerisation would be even slower (5).

Recently, an alternative approach to crosslinking methacrylated dextran polymers into a hydrogel is has been developed, using radiation instead of the APS TEMED system. This approach circumvents the use of toxic crosslinkers. If the methacrylated dextran monomers are pure, the crosslinked hydrogel is devoid of unreacted small molecules and no further purification is necessary. Additionally, if a high enough radiation dose is applied, typically 25 kGy, the gel is sterilized simultaneously allowing immediate seeding. Crosslinking by radiation is therefore a very promising option for dextran hydrogel synthesis on an industrial scale (7).

Hydrogel leakage test

1H nuclear magnetic resonance spectroscopy has been used to verify that during storage in demineralized water, no dextran or other starting products leak from the hydrogel. Trace amounts of organic residue were detected but could not be assigned. Besides the solvent peak, there is no overlap between the spectra of DMAP, GMA and that of the storage water as can be seen below.

Figure 1 The H-NMR spectra of DMAP (top), GMA (middle) and the storage water (bottom) do not show any overlap. This shows that the final dextran gel does not contain any toxic DMAP or GMA.

Waste production

Hydrogel synthesis

Only aqueous waste is produced during the synthesis of methacrylated dextran and the polymerisation into a hydrogel. Aqueous waste is water with small amounts of organic molecules dissolved in it. In our case these are organic molecules are the solvent DMSO, DMAP, unreacted GMA and methacrylated dextran, NaOH, HCl, synthesis byproducts and byproducts formed during the radical polymerisation (including unreacted methacrylated dextran). None of the waste products can be recycled due to dilution or contamination with other compounds.

Especially during dialysis, a large amount of aqueous waste is generated. Approximately 50 liters of aqueous waste is produced for 80 mL of hydrogel. The amount of waste can be drastically reduced by changing using a smaller dialysis volume. We used a dialysis volume of 4 litres to accelerate the dialysis. If a smaller volume would be used and equilibrium between the dialysis bag and the medium is established each time, we expect that we can halve the amount of waste, at the cost of a longer purification duration.

After use

The dextran-GMA hydrogel is biodegradable and is therefore safe to use in ecological environments. The bacteria are equipped with a kill switch which will also kill the bacteria after the patch has served its function and is thrown away. In this fashion waste production is kept at a minimum.

Costs

We are using a dextran hydrogel for our system. Medical-grade dextran is expensive. However, the production of one hydrogel only requires small quantities (100 μg). Therefore, the estimated dextran costs for making one 400 μL hydrogel in patch form is around €0,65. Adding other reagents will bring the total costs to approximately €1,00 for one hydrogel in small scale conditions. These costs can be optimized by upscaling. Modifying the bacteria is a one-time expenditure. Afterwards they can be cultured and re-used for a long period. Culturing the bacteria can be performed on a large scale. Costs of medium are very low. Overall, our living material can be produced very cost-efficiently and upscaling would further reduce cost. Before seeding our hydrogel with bacteria, they can be modified for expression of different proteins. By also culturing these bacteria that express infection-reducing enzymes, patches that fight infections can be created by a very modest increase in costs.

Conclusion

Although we use toxic chemicals for the synthesis of the hydrogel, they are removed by extensive dialysis of the methacrylated dextran resulting in a biocompatible end-result. Furthermore, a leakage test showed that no leakage occurs from the gels over a period of two weeks. This means that the chemicals do not yield any risk for users of the bandage. The use of a novel, radiation-based crosslinking method and alternative catalysts could reduce the toxicity of the synthesis and polymerization reactions, which would be beneficial for manufacturers. The gel is biodegradable, and its production is relatively cheap.

To address this issue, we consulted several experts from the field of wound healing, diabetic wounds and burn wounds. The enzymes that will be secreted by our patch need to be safe in human systems. Lysostaphin is one of these enzymes. Lysostaphin efficacy has been extensively tested in murine and rabbit models, confirming its effectiveness against S. aureus causing several types of infections (e.g. nasal, orthopaedic implants, abscess lesions, aortic valve endocarditis) (9–11). In those cases, lysostaphin treatment was reported to be non-toxic (11). Clinical studies must be performed to confirm these statements. Therefore, clinical trials are necessary to show that lysostaphin causes no harm when it enters the body through the wound. Although most bacteriocins have no effect on human cells in low concentrations, less is known about higher concentrations. Besides lysostaphin, our E. coli will produce pyocin, a bacteriocin against Pseudomonas aeruginosa. Pyocin has been shown to be effective against P. aeruginosa lung infections in mice, with its required dose 100-fold lower than the commonly used antibiotic tobramycin (12). It still needs to be determined whether it is safe in humans, but there is no indication it will be.

The hydrogel that we use is biocompatible and has been used in humans before. However, GMA that is necessary for the synthesis is highly toxic and carcinogenic. Dialyzing the methacrylated dextran product for one week ensures complete removal of GMA. 1H nuclear magnetic resonance spectroscopy has been used to verify that during storage in demineralized water, no dextran or other starting products leak from the hydrogel . Therefore we know it is safe to use for an application on the skin. The trace amounts of organic material that do leak out could not be assigned to any known reagents or (by)products but it was collected from the 50 mL of storage water. This means that the presence of organic materials in the storage water can be neglected as their concentration is extremely low.

The strains of E. coli (BL21(DE3) and BLR) that we use do not present a direct threat to human health, plants or microorganisms, since it is an biological agent of category 1 (13–15). This means that it is very unlikely that they are harmful to humans. Besides the bacteriocins that are secreted, E. coli has been modified to express a number of different proteins. As mentioned above, these proteins are harmless.

In conclusion, to make our gel user friendly and safe, we used a hydrogel that is not hazardous and is safe when in contact with living organisms. Our E. coli strain is harmless. However, it must still be confirmed that the used bacteriocins are safe to humans. To this date there is no indication that it the bacteriocins would be any more hazardous than conventional antibiotics.

Good working conditions are conditions which ensure that people can work healthily and safely (6). Therefore, it is important to make an inventory of the possible existing hazards. Thinking about our “hydrogel patch”, possible hazards are the dextran material, the diffusion of lysostaphin and pyocin and the bacteria that are inside the gel. As we discussed above under patient safety, both of theseboth molecules seem safe to humans. The same applies to the E. coli strain used. Obviously, it remains important to wash hands before and after handling the patch or touching the wound.

Moreover, to work healthily and safely, is it also important that the gel can be applied easily. Therefore, it should not require much physical work and should not be harmful for the nurses. It would be beneficial if the plaster looks normal, something that feels familiar to the patient and nurse. It should not give them the idea that there are living modified bacteria inside the plaster itself to prevent scaring patients or medical professionals.

From our conversations with medical professionals, we learned that the plaster will need to be replaced regularly. To prevent pain or inconvenience to the patient or nurse, the plaster must thus be easy to apply and to remove. As the plaster will be delivered in a sterile package ready for application, the application should be straightforward and easy. This requirement is thus met. Overall, it can be concluded that our plaster will not make the work of nurses or doctors any more difficult or dangerous than it currently is.

Several experts and stakeholders asked us whether the risk exists that foreign bacteria infiltrate Gelcatraz. Infiltration would form a serious safety hazard. What if malignant bacteria colonize the material and use it to secrete toxins to patients? We thus researched this issue and consulted several experts on the topic.

In principle, we do not think that the risk of infiltration is high, as ‘our’ E. coli bacteria secrete toxins. This does not make the material an attractive environment for foreign bacteria, or at least the targeted ones. Once we will have incorporated a broad spectrum of toxins to be secreted on demand, the risk of infiltration by malignant bacteria will be minimal.

What remains, is the risk of non-targeted bacteria infiltrating the gel. Think of benign bacteria present on our skin, which will be attracted by the nutrient-rich environment of the gel. Although this will not pose a safety-hazard, it would be problematic when they outcompete E. coli and consume the nutrients. It is, however, more likely that E. coli remain the dominant species, as they are present in large amounts. As foreign bacteria lack the binding affinity for Dextran that our E. coli have, this will only happen when the gel forms a sufficiently attractive environment and when they can live together symbiotically with E. coli. This is not necessarily a bad thing, such as in the case of mutualism or commensalism. Whether this happens, will need to be tested experimentally. In case it happens, we researched some preventive measures.

Changing the plaster

A simple measure may be frequent replacement of the plaster. This would ensure fresh nutrients and fresh E. coli bacteria. A disadvantage of this is inconvenience for the patient, which was highlighted by doctors from the Maasstad Hospital. Clinical studies will need to show the optimal treatment duration and replacement frequency.

Quorum Sensing

Bacteria use quorum sensing to regulate activities on a larger scale. This is mostly intraspecies but the molecule AI-2 (Autoinducer 2) has been suggested to be a universal quorum sensing molecule, allowing inter-species communication. Production of this molecule has been found in a wide range of both gram-positive and gram-negative species, by means of the protein LuxS (11). By implementing a luxS knockout in our own bacteria, combined with an AI-sensing gene we can induce production of bactericides upon invasion of unwanted bacteria into our living material. A similar mechanism has also been attempted by NTU-Singapore 2008 iGEM team.

Silver coating

Many metals have antibacterial properties (oligodynamic effect), but silver is the most effective and the least toxic to humans (16). It is commonly used in medical treatments. Silver ion release from silver composites has also been presented as an effect antimicrobial method (17).

However, silver has both an effect on S. aureus and E. coli. Gram-negative bacteria such as E.coli suffer more structural damage from silver than Gram-positive bacteria like S. aureus (18).

During the European meetup in Munich we explained iGEM Düsseldorf how we want to make a living material. They recognized the problem of bacterial overcrowding within our construct. Since our bacteria will keep dividing and cannot escape the hydrogel, at some point bacteria will be over-saturated which can lead to problems in protein expression or even dead. They offered to help us, and we had a meeting with them over skype. They were making a construct that uses the Lux Quorum sensing. Their construct contained 3 plasmids: Luxl (synthase for AHL), LuxR (molecule where AHL binds and can then bind to promotor) and pLux (promotor to which LuxR +AHL can bind). When the promoter gets activated, a phage lysis gene will be synthesized. In our case they offered to send us the plasmids without the Luxl in the construct, so when we add AHL to the hydrogel medium, cell growth will be inhibited. This is a great solution to keep our bacteria in control, as we want them to stay in steady numbers within our hydrogel, with steady lysostaphin production.