>

Design

Project design

Colorectal cancer is one of the most prominent and deadly cancers [1]. Existing treatments can cause severe side effects and the scope of available therapies is limited. In an attempt to expand the list of available treatments we focused on a non-invasive way of treating colorectal cancer. We also aimed to reduce the side effects of anti-cancer treatments by a targeted localisation of our deliverable drug combined with regulated secretion of an anti-cancer peptide.

Repressilator system with a three gene — Figure 1: Final design of our “B.O.T” bacteria. Probiotic E. coli Nissle 1917 contains two plasmids. A repressilator plasmid containing the repressors necessary to produce oscillations and reporter genes for visualizing the oscillations. The plasmid also contains half of the kill-switch system implemented for controlling bacterial survival. The sponge plasmid is necessary to make the oscillations robust. It also codes for the anti-cancer peptide azurin. Finally, it also contains the other half of the kill-switch system, together they act as selective pressure to keep both plasmids in the cell.

Here we present B.O.T, an engineered probiotic containing an oscillatory circuit called the “repressilator” [2] [3] that secrete an anti-cancer peptide (azurin) at regular intervals. Our goal is to have the probiotic bacteria ingested orally in the form of a pill. The bacteria would be released in the patient’s gut. Our B.O.T. strain would then naturally migrate and colonize tumor-like environments in the gut and secrete the anticancer peptide in an oscillatory manner. After a few weeks the natural turnover of the gut would lead to the patient having to take another pill.

Engineering E. coli Nissle 1917 with the repressilator system

First, we introduced the repressilator circuit [2] [4] into E. coli Nissle 1917. The system is composed of two plasmids: the repressilator plasmid and a sponge plasmid. The sponge plasmid is required to reduce the noise of oscillations. We tested the original repressilator system in our probiotic strain and measured the fluorescence of the different reporter genes over time, (see Figure 2). We showed that the fluorescence proteins from the repressilator plasmid were expressed in an oscillatory manner like in the original paper. We then went on to edit the sponge plasmid and inserted either a reporter gene or the gene encoding for our anticancer peptide downstream of the PLtetO1 promoter (Figure 2).

If azurin was put under the regulatory control of P_tetR, with every peak in P_LtetO1, we would have the production and secretion of the anti-cancer peptide. We used modelling of the repressilator to predict the oscillatory pattern of the different reporter genes over time, if you want to learn more visit the modeling page.

Testing azurin production on colorectal cancer cells

We wanted to test azurin production by our probiotic E. coli Nissle 1917 and its effect on the Caco-2 cell line, a common cell line for colorectal cancer research. Since azurin was going to be produced and secreted by our bacteria, we designed experiments to first establish the Caco-2 cell line response to existing cancer drugs (Anisomycin, Cisplatin, Doxorubicin, Salirasib and TNFa) with an MTT assay. This survival essay served as a baseline control for us to later test the higher or lower effect of azurin on those cancerous cells. In the end we found azurin to have no effect on the Caco-2 cell line. If we had had more time we would have liked to test the effect of azurin on the HPK 116 colorectal cancer cell line. In parallel, we tested the expression and secretion of azurin by our probiotic bacteria. The expression of azurin was not as straightforward as expected and we had to go through multiple iterations of our plasmid constructs to obtain purified azurin from the cell lysate where we expected azurin to be exported to. If you want to go into more detail visit our azurin purification page. The next step would have been to grow azurin-expressing bacteria directly on the HPK 116 colorectal cell line to determine the efficiency of azurin production and secretion and the ability of our probiotic bacteria to kill cancerous cells, but we unfortunately did not manage to get to this point experimentally.

Establishment of a kill switch

The concern for biosafety is a top priority, so we therefore wanted to establish a system to make our project safe (if you want to learn more about our human practices safety work we invite you to read the GMO survey). Questions like “how could we control the survival and localisation ofour bacteria?” and “How could we avoid them from disseminating into the environment?” arose. Consequently, we wanted to implement a controllable kill switch in our system. In parallel to the cloning of the kill switch in the wet lab we used a model to help us predict the survival outcomes for the various parameters we wanted to test (if you want to learn more about the kill-switch model, please visit our modeling page). The final designs of our plasmids contain specific temperature and phosphate level sensitive promoters that would lead to cell death by unbalancing the levels of toxin and anti-toxin present in the bacteria if not maintained at ~37 C or in the correct phosphate concentrations. The temperature sensitivity of the kill-switch system allows it to respond differentially when the bacteria are in or out of the body. The phosphate sensitivity allows the kill-switch to respond differentially when the bacteria are in the colon or in the blood, which have two very different levels of phosphate in the environment. The toxin/anti-toxin couples expressed by the kill-switch we tested were ccdB/ccdA and miniColicin/IM2. The kill-switch would also act as a plasmid retention system since the toxin lingers a while and degrades slower than the anti-toxin in the organism, the loss of the plasmids would therefore lead to the death of the cell. This is necessary as the two plasmids required for the repressilator could lead to a heavy metabolic load and affect the fitness of our engineered strain.

Table 1: Table of conditions and expression of the kill switch

	Colon	Blood circulation	Phosphate pill to colon	Outside of body
[Phosphate]	Low	High (1mM)	High	Low
Temperature	37°C	37°C	37°c	<25°C
Toxin expression	Low	Low	Low	High
Antitoxin expression	High	Low	Low	Low
*Status of E. coli* Nissle 1917**	Alive	Dead	Dead	Dead

With the conditions imposed by either the promoters or RNA thermometers (non-coding RNA strands that regulates gene expression depending on the temperature), we could modulate the expression of the toxin and the anti-toxin, which leads to a controlled growth or death of our bacterial cells.

As this design was relatively complex, we tested out the toxin-antitoxin combinations using single plasmids. If you want to know more about the kill switch, click here.

References

[1] “Colorectal Cancer Statistics | How Common Is Colorectal Cancer?” n.d. Accessed October 26, 2020. https://www.cancer.org/cancer/colon-rectal-cancer/about/key-statistics.html.

[2] Elowitz, Michael B., and Stanislas Leibler. 2000. “A Synthetic Oscillatory Network of Transcriptional Regulators.” Nature 403 (6767): 335–38. https://doi.org/10.1038/35002125.

[3] Potvin-Trottier, Laurent, Nathan D. Lord, Glenn Vinnicombe, and Johan Paulsson. 2016. “Synchronous Long-Term Oscillations in a Synthetic Gene Circuit.” Nature 538 (7626): 514–17. https://doi.org/10.1038/nature19841.

[4] Riglar, David T., David L. Richmond, Laurent Potvin-Trottier, Andrew A. Verdegaal, Alexander D. Naydich, Somenath Bakshi, Emanuele Leoncini, Lorena G. Lyon, Johan Paulsson, and Pamela A. Silver. 2019. “Bacterial Variability in the Mammalian Gut Captured by a Single-Cell Synthetic Oscillator.” Nature Communications 10 (1): 4665. https://doi.org/10.1038/s41467-019-12638-z.