Team:UNILausanne/Proof Of Concept
Proof of concept
As a proof of concept, we used different methods to test the expression of azurin in an oscillatory manner. Moreover, we characterized the effect of anti-cancer drugs and azurin on colorectal cells.
Expression of azurin in an oscillatory manner
For our project, we wanted to express a cytotoxic protein, azurin, in an oscillatory manner. In this way, it will be delivered during the night in colorectal tumors to improve its anti-cancer effect and will act as a chronotherapeutic drug. In order to achieve this, we assessed the characteristics of the oscillatory network and evaluated the expression of azurin. Finally, we investigated the combination of both parts of the project.
To combine both parts, we added azurin to the sponge plasmid that is part of the repressilator system. This choice made the combination of both parts easier, as we did not have to modify the core repressilator plasmid. To track the oscillatory expression of azurin over time, we asked the Protein Analysis Facility of UNIL (PAF UNIL) to perform a Western blot for the detection of azurin in our samples at different time points.
The samples were prepared by inoculating E. coli Nissle 1917 Δclb transformed with:
1- BBa_3482024 + BBa_3482038 = plasmid expressing the truncated azurin (BBa_2500001) with a PeIB-5D secretion tag and 3XFlag tag under the control of a pTet promoter(pLTP145) as a positive control for constitutive expression of azurin.
2- BBa_3482023 + BBa_3482024 = repressilator(pLTP234) and sponge plasmid(pLTP145) without azurin as a negative control
3- BBa_3482023 + BBa_3482024 - BBA_3482038 = repressilator(pLTP234) and sponge plasmid(pLTP145) expressing truncated azurin with a PeIB-5D secretion tag and 3XFlag tag as our tested oscillatory expression.
Bacteria were incubated for 4 h in LB with 50 ng/mL aTc to synchronize the repressilator system, then pelleted and resuspended in fresh LB medium without aTc to remove inhibition of the repressilator, thus enabling oscillations to start. OD600 was measured every hour and dilutions were made to maintain the cells in exponential phase without exceeding OD600= 1.0. Samples were taken at each time point for a final OD600 = 1.0. Cells were centrifuged, the supernatants were taken and precipitated. The pellets were resuspended in lysis buffer and lysed by lysozyme 1X treatment (incubated for 30 min) and sonication (20 sec in 1 sec intervals at 10% amplitude using MS73 probe). Samples were then sent to the PAF for immunoblotting analysis.
In total, four time points (T0: 0 h, T1: 2 h, T2: 5 h and T3: 7 h) in duplicates (#1 and #2) were evaluated. A Western blot was performed using a primary 3XFlag tag-specific antibody (Flag M2 (Sigma F1804)), plus a secondary anti-azurin antibody (azurin biobyt orb 379723) (Figure 1A).
In addition, a nonspecific secondary antibody was used to assess the total protein following the Replex method (Figure 1B). This allows us to normalize our quantification and thus to have more reliable data analysis.
In our Western blot we expected the positive control to have a band of similar intensity at each time point in the pellet as well as in the supernatant, because azurin is constitutively expressed and secreted. For the oscillatory sample, we were expecting a band of changing intensity for the different timepoints in the pellet as well as in the supernatant. For the azurin antibody, we expected to see one band at the same size as for the samples where we used the FLAG antibody. Finally, for the negative control we did not expect any band.
As we can see in Figure 1A, the duplicates of the positive controls are showing bands at each time point. However, the intensity is higher at T0 than for the other time points. This difference persists when normalized for the amount of total protein (Figure 2). We hypothesize that this difference comes from the fact that cells at T0 were in the presence of aTc, while the rest were not for T1-T3. Moreover, we noticed a non-expected band in one sample (indicated by the red star), that we could not explain. We can observe a band at T0 and the absence of the tagged protein from T1 to T3 in each tested oscillatory expression sample (pellet and supernatant). This might correlate to the decrease of expression due to the oscillations. However, as the decrease is also present in the positive control, the experiment would have to be repeated such that T0 is also taken from cells in an exponential phase. Moreover, we do not see an increase in the oscillatory sample that we would expect after 7 h. An experiment in the future would have to be repeated for longer than 7 h to see if we can detect an increase at later time points.
Unexpectedly, the anti-azurin antibodies column shows multiple bands, this may be protein aggregation. Finally, the negative control does not show any band as expected.
In summary, we established an experimental setup that should allow us to detect the oscillatory release of azurin and we performed a preliminary experiment. In the future, we will repeat this experiment under optimized conditions and hopefully detect oscillations in azurin production.
The effect of azurin on colorectal cancer cells
To assess the effect of azurin on cancer cells, we performed a cell viability assay using the MTT (bromure de 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium) method. To do this, we chose a well-documented colorectal cancer cell line as our model: Caco-2 cells. Indeed, it is a largely used cell line which allows us to compare our results of the MTT assay with the literature. The Caco-2 seems to be more resistant to current treatment than other types of cancer cell line. This resistance might come from genotypic variation as it has been shown that this cell line can show phenotypic variation [1, p. 2], Thus, we first characterized the proliferation of our Caco-2 cell line in the presence of cytotoxic drugs. In a second step, we administered them with bacterial extracts from transformed E. coli Nissle 1917 Δclb either containing or lacking the azurin coding gene.
Assessing drug effect on our colon cancer model cell line (Caco-2)
To compare the efficiency of azurin with current treatments, we first conducted preliminary experiments on our cancer cell line. We used an MTT assay to evaluate the viability of our cells 24h after treatment. We selected a wide range of cytotoxic chemicals commonly used in labs or as chemotherapeutic agents.
Caco-2 cells were seeded at ~12’500 cell/mL in modified DMEM media with 10% fetal bovine serum (FBS), 1X non-essential amino acids, and 1X penicillin / streptomycin in 96 well plates. Chemotherapeutic compound stocks were diluted to desired concentrations in the same medium and then applied to the cells. Cell viability was measured after 24 hours using the MTT assay method.
We can observe in Figure 2 the effect of cytotoxic agents on the Caco-2 cell model. Notably, we observe a lack of Lethal Dosage for 50% cell death (LD50) for the tested chemicals. This suggests a lack of sensitivity of the cells to cytotoxic compounds as well as showing the variability of response of the cells to chemical agents. We can see that the greatest effect can be found using doxorubicin at 100 uM. Thus, the LD50 of this DNA damaging agent is probably between 10 and 100 uM. With this assumption, we will use doxorubicin as a chemotherapeutic chemical of reference to characterize the effect of azurin. We confirm the lethality during our assay by using 2 positives control acting on the cells by membrane destabilisation.
In summary, this experiment helped us understand the behavior of Caco-2 cells in presence of cytotoxic compounds and the intrinsic sensibility of our cells.
Assessing bacterial extract effect on our colon cancer model cell line (Caco-2)
Finally, we tested the effect of bacterial samples (either supernatant or lysate) of E. coli Nissle 1917Δclb over time to assess the potential chronotherapeutic effect of the repressilator coupled with azurin expression.
Bacteria were incubated overnight in LB with 50 ng/mL aTc to synchronize the repressilator system. Cells were pelleted and then resuspended in fresh LB medium without aTc to release the inhibition of the repressilator, thus enabling oscillations to begin. OD600 was measured over time and dilutions were made to maintain cultures in exponential phase. Samples were taken periodically and OD600 normalized to 0.5. Upon centrifugation, supernatants were collected and pellets were resuspended in lysis buffer and lysed by lysozyme , then sonication, following the same procedures as described earlier.
We can observe no decrease in cell viability of 50% or less (Figure 3.A and 3.B). In consequence, no LD50 can be defined from our samples. We expected a lower viability at the first and last time point and a higher viability in between, as it would have been correlated to an oscillatory expression of azurin. Interestingly, we can see in Figure 3.B, the repressilator and sponge plasmid with azurin sample at 6h shows a drop in cell viability which could be due to a peak in oscillation in the expressed azurin under the repressilator. Lastly, we cannot see a cytotoxic effect of azurin on our Caco-2 cell line with the administered concentration.
In the light of these results, we explored the literature further and discovered that although largely used as a colorectal cancer model, Caco-2 cells are primarily used as a model for intestinal absorption. In fact, Caco-2 cells are commonly used as such due to their unique property to replicate the intestinal barrier in vitro[2]. In addition, we found that they possess a mutated p53 pathway, thus being categorized as p53 negative cancer cell line. Although not totally immune, reduced sensitivity to azurin can be expected, given the absence of its main molecular target [3].
We propose for future iGEM teams willing to characterize intestinal absorption of compounds in humans to use the Caco-2 adenocarcinoma model cell line. Indeed, we have established a protocol for Caco-2 cell culture and for a viability experiment. In addition, we characterized the viability of these cells treated with several cytotoxic compounds.
References (click to see)
- [1] T. Lea, "Caco-2 Cell Line", in The Impact of Food Bioactives on Health: in vitro and ex vivo models, Éd. Cham: Springer International Publishing, 2015, p. 103‑111.
- [2] I. D. Angelis et L. Turco, "Caco-2 Cells as a Model for Intestinal Absorption", Curr. Protoc. Toxicol., vol. 47, no 1, p. 20.6.1-20.6.15, 2011.
- [3] H. M. Barzelighi, B. N. Esfahani, B. Bakhshi, B. Daraei, S. Moghim, et H. Fazeli, « Influence of Heterologously Expressed azurin from Pseudomonas aeruginosa on the Adhesion and Invasion of Pathogenic Bacteria to the Caco-2 Cell Line », Probiotics Antimicrob. Proteins, vol. 12, no 2, p. 697‑704, juin 2020.
