In this study, we observed that human lung cells transfected with DcsE-FLAG-GFP are capable of synthesizing functional DcsE, which catalyzes the first step towards D-CS synthesis. D-CS is an optimal candidate for a genetic chemoprophylaxis against TB because it is the shortest known pathway for an anti-TB antibiotic with biologically available precursors (Table 1). DcsE tagged with FLAG-GFP can be expressed in A549 cells (Fig. 2) and purified DcsE (Fig. 3) catalyzes the formation of L-OAS from L-serine and acetyl-CoA (Fig. 5). L-OAS produced by DcsE has the expected retention time of 1.1 minutes (Fig. 5a), a mass of 148.13 m/z (Fig. 5d), and produced the same 106 m/z and 131 m/z fragments at 5V (Fig. 5e) as the 0.5mM L-OAS standard in reaction buffer (Fig. 4e). No L-OAS was detected from a reaction containing purified FLAG-GFP (Fig. 5c).
Production of an active DcsE enzyme in human lung cells demonstrates progress towards a genetic chemoprophylaxis against TB. Previous studies have shown that other TB antibiotics such as isoniazid can be used prophylactically to prevent TB infection in those infected with HIV and people who have tested positive for TB by Tuberculosis skin test15. Isoniazid is not the only TB antibiotic used chemoprophylactically; rifampicin, pyridoxine, and combinations of rifampicin and pyridoxine have also been shown to reduce the incidence of TB infection compared to a placebo16.
One limitation to using D-CS for the proposed gene therapy is the comparatively high minimal inhibitory concentration (MIC) for Mtb of 10–50µg/mL which could restrict the efficacy of the therapy depending on how much D-CS can be produced in humans. In this study, 40.3 μM L-OAS was produced after a 1 hour reaction at 30ºC which equates to 5.94 μg/mL L-OAS, meaning the concentration of a pre-cursor is lower than the MIC of D-CS. Future experiments in conditions that more closely model the complex environment of the human system could better determine realistic concentrations of D-CS that could be produced. However, typically, D-CS does not reach MIC concentrations in patients, suggesting reaching MIC is not required for therapeutic efficacy. D-CS is known for its poor lung cavity penetration; the clinical dose for adults of 250mg only reaches MIC concentrations of less than 2 μg/mL in the lung17 meaning that small amounts of D-CS being produced intracellularly while infection is mild could be clinically relevant.
To apply the chemoprophylaxis of TB using antibiotics genetically, there are multiple steps that need to be taken to ensure safety. Moving forward, we propose developing an excisable gene therapy that delivers D-CS in the presence of TB infection (Fig. 6a). The site specific recombination of the CRE/loxP1 system could be used to permanently excise dcsABCDEG in the case of adverse effects. As visualized in Fig. 6a, expression of the recombinase CRE is activated by the drug tamoxifien18. When expressed, CRE cleaves the two loxP sites thus excising the construct responsible for production of D-CS (Fig. 6b). The CRE/loxP1 system is a suitable system for this application because it is precise, controlled by a pharmaceutical that is not naturally present in human cells, and is able to quickly and permanently stop D-CS expression. Additionally, in the absence of infection, dcsABCDEG will not be expressed due to the infection responsive promoter element (InfRE) and D-CS will not be produced (Fig. 6c). D-CS must only be produced in the presence of M. tuberculosis to reduce unnecessary production of the antibiotic (Fig. 6d). To accomplish this, a protease produced by M. tuberculosis could be used to control the expression of functional enzymes. In this model, the sequence SLKPASAGGG (represented by dotted lines) cleaved by MycP119 could be between each of the six enzymes (Fig. 6). When Mtb and one of its MycP1 protease are present, the enzymes would be expressed and cleaved from each other thus becoming functional. By including an infection induced promoter, MycP1 sites between each enzyme to control functionality of enzymes in the presence of TB, and an excisable plasmid controlled by CRE/loxP1 in our design, the expression of D-CS synthesis genes is controlled.
The system used to deliver the genetic chemoprophylaxis must be long lasting and well tolerated. Adeno-associated viruses (AAV) are one way to deliver a construct similar to the one proposed in Fig. 6a. AAV are a versatile and effective way of delivering genetic material via a protein shell containing single stranded DNA20. The genome capacity of AAV is generally considered to be around 4.5-5kb which is below the estimated size of the plasmid proposed in Fig. 6. However, AAV5, a serotype of AAV, has been shown to incorporate genome sizes up to 8.9kb21 which is larger than the plasmid proposed in Fig. 6. Cationic liposomes (CLs) are another example of a technology capable of delivering a gene therapy. CLs are non-viral and use a closed lipid bilayer membrane to deliver genes and protect the DNA from degredation22. CLs are very customizable and can transfer up to 1,000kb22 into cells making them another promising technology for gene therapy applications. Overall, both AAV and CLs are delivery methods that could be used to deliver the construct tissue specifically, safely, and dependably.
By observing heterologous production of functional DcsE-FLAG-GFP, we provide evidence that the first step for prophylactic D-CS synthesis is possible in human cells. Our next steps include developing a method to observe this enzymatic activity in live A549 cells and eventually advance to synthesizing all six enzymes in the D-CS pathway via the proposed plasmid. Overall, with the rapid growth of cell based therapies, we seek to create the tools now to prepare for a future state where gene therapies are commonplace, safe, economically plausible, and well accepted.