We sought to develop a new protein that can successfully neutralize all four RSPOs in order to antagonize RSPO in malignancies. This protein, as compared to the RSPO3-specific antibody, has the therapeutic potential to treat both colorectal cancers caused by RSPO2 and RSPO3, as well as other cancers caused by RSPO1 or RSPO4. The ECD domains of LGR4 and ZNRF3 were joined to form a chimeric protein that was designed and engineered based on earlier biochemical and structural biological research (Figure 1a)^1-4. In order to connect these two domains and offer adequate flexibility and solubility, a linker region was optimized in between (Figure 1b). We examined several signaling peptides from various transmembrane proteins in search of the best secretion capacity. The results demonstrated that the signaling peptide from Frizzled5, a Wnt receptor, can most efficiently induce the chimeric protein's secretion into the conditioned media (Figure 1c). By joining the IgG Fc (human IgGH3 Fc) region to the carboxyl terminus of the chimeric protein, further stability was provided to the protein (Figure 1d). It was demonstrated experimentally that Fc-fused chimeric protein could persist for over 4 weeks at 37 Celsius without degrading or aggregating in FBS-containing medium, which persists noticeably longer than the proteins without (Figure 1e).

Also, a V5-tag was engineered at the animo terminal of the protein as an epitope for easy detection. Thus, we gave the chimeric protein's final iteration the name RTAC (RSPO-Targeting Anti-cancer Chimeric protein) (Figure 1d).

The next goal was to characterize RTAC's capacity to bind to RSPOs. First, we transiently transfected RTAC and four HA-tagged RSPOs into HEK293 cells. After 48 hours, we collected secreted RTAC and four RSPOs from the conditioned medium. We then co-incubated RTAC with individual RSPO in vitro, respectively, and used α-HA antibody-conjugated agarose to pull down the RSPOs. The results showed RTAC exhibited pan-RSPO binding capacity towards all RSPOs, which was stronger than LGR4 or ZNRF3 ECD alone (Figure 2a,2b).

Following the binding studies, we attempted to evaluate RTAC's RSPO-inhibiting potential. TOPFlash luciferase is a widely used assay to evaluate Wnt/β-catenin signaling strength. Renilla luciferase is used as an internal control while the firefly luciferase is engineered downstream of the TCF/LEF response element to offer specific and strong Wnt/β-catenin signal output (Figure 3a).

The plasmids expressing these two luciferases were transiently transfected into HEK293 cell before further experiments. We observed in these HEK293 cells that the addition of RSPO3-containing conditioned medium could augment the signaling intensity by multiple folds. It was satisfactory to see that RTAC treatment could effectively neutralize the Wnt/β-catenin activating effect of RSPO3 (Figure 3b).

Besides the luciferase assay, multiple experiments were exercised to confirm the anti-RSPO function of RTAC. Treatment of RTAC effectively nullified the accumulation of cytosolic β-catenin and Wnt target gene expressions caused by additional RSPO3, shown by Western Blotting and quantitative RT-PCR assay, respectively (Figure 3c, 3d).

These findings were subsequently confirmed in colorectal cancer cell RKO, where the introduction of RTAC greatly negated the effect of RSPO. (Figure 3e-3g).

It is commonly known that Wnt/β-catenin signaling activation is crucial for the growth of many tumor cells. For this reason, we selected several Wnt-sensitive cell lines, such as RKO, Patu8988s, and U2OS to investigate the broad effectiveness of RTAC in inhibiting tumor cell proliferation. RSPO3-containing conditioned medium was added into the cell culture every two days starting from Day0. And purified RTAC after size-exclusion chromatography was added together with RSPO3. MTT cell growth assay indicated that RTAC addition could significantly restrict the growth of these cells at the presence of RSPO (Figure 3h). And higher dose of RTAC showed stronger inhibitory effect in RKO cell with statistical significance (Figure 3h). These findings present conclusively that RTAC blocks in vitro tumor cell proliferation and RSPO-mediated Wnt signaling. Thus, our engineering of RTAC came to success.

We next aimed to employ engineered microorganism to create a self-tunable machinery to deliver RTAC to tumor cells. Because of the existence of microbiota in the intestinal tract, we proposed that this microorganism can be used as a probiotic to prevent or treat RSPO-hyperactivated colorectal cancer. Brewer’s yeast (Saccharomyces Cerevisiae) is a model organism that has been studied for decades, which genomic information is clear. It is also highly biosafe to human as it can behave as a beneficial or commensal microorganism in human guts⁵. Meanwhile, this eukaryotic system can allow mature and functional RTAC to be expressed and secreted. Because of these advantages, we decided to employ S. Cerevisiae as our vessel to deliver RTAC to the intestinal tumor.

Based on the previous report, we decided to reprogram the yeast’s mating pathway to allow it to sense tumor-environmental specific molecules^6,7. Extracellular ATP (eATP) was reported to have specifically strong presence at the tumor tissue up to hundreds of uM concentration⁸. Thus, this molecule serves as an ideal signal to trigger tumor-tissue specific expression of RTAC. Following the previous research⁹, we first reprogrammed the mating pathway by removing the receptor ste2 through gene editing to ensure the pheromone α can no longer stimulate the downstream pathway. Also, the mating inhibiting factor sst2 was removed to intensify the signal output (Figure 4a). This double-knocked out yeast strand was used in our further experiments.

Meanwhile, we introduced modified human P2Y2 receptor and chimeric Gα as the previous literature reported⁹. The introduction of these two components will enable the yeast to sense eATP and transduce the signal to the native MAPK signaling in the yeast^6,7,10(Figure 4a). Thus, in the absence of ste2, the yeast’s mating pathway could only be activated through eATP-P2Y2-Gα-MAPK axis. To minimize the usage of selection marker, we linked P2Y2 receptor and the chimeric Gα in the same vector with a P2A peptide sequence in between. The P2A peptide can undergo a translational recoding event to generate two independent proteins before and after this “break”¹¹. By fusing P2Y2 receptor and a common fluorescent protein mCherry, we evaluated the efficiency of the P2A peptide in our system. First, from Western Blotting assay, we observed that the HA-tagged P2Y2 receptor displayed as approximately 40kDa on the gel, which was its native molecular weight (Figure 4b). If P2A failed to function, the presumable P2Y2-mCherry fusion protein had a MW around 78kDa, which did not appear on the gel. Also, the fluorescence of mCherry confirmed the cleavage efficiency of the P2A peptide (Figure 4c).

Figure 4b: WB results confirming the self-cleavage function of P2A peptide in P2Y2-P2A-mCherry and P2Y2-P2A-Gα plasmids

Figure 4c: Fluorescent microscopy results showing mCherry localizes in the cytosol instead of on the membranes

As a soluble fluorescent protein, mCherry localizes in the entire cytosol. If the P2A peptide was unfunctional, P2Y2 could tether mCherry to the membrane. In Figure 4c, it was very clear to visualize that the fluorescence of mCherry spread out within the whole yeast cells without recognizable membrane residence, which demonstrated that mCherry was released from P2Y2 receptor. Indeed, when we transformed P2Y2-P2A-Gα plasmid into the yeast, we confirmed that P2Y2 and chimeric Gα were generated in the yeast separately by Western Blot (Figure 4b). To conclude, these results indicate that the eATP sensing parts are engineered into the yeast successfully and our P2A system can efficiently function.

After the upstream of the mating pathway was reprogrammed, we started for the downstream part. Since pFUS1 promoter target gene expression is the final output of the mating pathway^6,7, we constructed a plasmid with RTAC following the pFUS1 promoter (Figure 4a). To maximize the secretion efficiency, we adapted the signaling peptide of RTAC to that of α-factor (MFα1) particularly for yeast expression⁹. As a test, we also constructed the pFUS1-mCherry plasmid to detect the sensitivity of eATP treatment. We combinatorially transformed P2Y2-P2A-Gα and pFUS1-mCherry, and cultured the transformed yeast with selection markers for 48 hours. Afterwards, eATP was added to the culture at different doses and the red fluorescence was monitered. The result showed that starting at 25uM, the yeast started to show red fluorescence under the microscopy (Figure 4d). And the fluorescence intensified with increased eATP concentrations (Figure 4d).

This data serves as strong evidence to show that our engineered mating circuit can successfully sense the eATP from the environment and express substantial amount of our desired target. After this validation, we eventually transformed P2Y2-P2A-Gα and pFUS1-RTAC together into the engineered yeast. Upon eATP treatment, RTAC can be expressed and secreted into the culture in a dose-dependent manner as expected (Figure 4e). We thus named this eATP-sensitive RTAC-expressing yeast as R-yeast.

Due to the absence of animal use permission, we planned to test the function of R-yeast in vitro. We then co-cultured R-yeast with RKO cell to demonstrate the RSPO-inhibitory effect. First, from Western Blotting, we observed that the cytosolic β-catenin diminished in the co-cultured RKO cells upon the addition of 200uM eATP (Figure 4f). This data was further assisted by quantitative RT-PCR result, in which R-yeast significantly inhibited the Wnt-target genes expression (Figure 4g).

Cancer cell growth was also examined. Addition of RSPO3 could significantly enhance RKO growth, which could be reverted by R-yeast co-culturing with 200uM eATP in the culture medium shown by MTT growth assay (Figure 4h).

As a negative control, normal brewer’s yeast could not affect the growth of RKO. This data together has provided a proof-of-concept of our designed engineered yeast (Figure 4h).

To conclude, we successfully showed that R-yeast can sense eATP and express RTAC dose-dependently. And the expressed RTAC can functionally inhibit RSPO-induced Wnt/β-catenin signaling activation and corresponding cancer cell growth. In principle, the expression of RTAC in R-yeast remains silent in the absence of eATP and will auto-sense and switch on when coming across tumor tissue. This self-tunable yeast poses minimal threat to human health thus may serve as a probiotic to prevent and treat colon cancers.

The brewer’s yeast is not considered as a biohazardous or pathogenic microorganism in the environment in general. However, considering the possible leakage and unexpected hazards, we decided to add a self-killing switch to completely resolve this issue. After extensive readings, we realized that there was no well-recognized kill-switch in yeast from the literatures. So, we decided to design and construct one on our own. First, we picked an essential gene in yeast as our target. By manipulating the expression of this gene, the survival of the yeast can be controlled. HSF1 was reported to be an essential gene in yeast, which pivotally governs stress response and other crucial processes¹², was chosen as our target. Meanwhile, we investigated the difference between the environment and the intestinal tract through readings and consulting. We realized that copper ion was abundant in the intestine but scarce in the environment¹³. Interestingly, the copper ion can activate copper-ion dependent promoter CUP1 for target gene expression^14,15. As a result, we replaced the original promoter of HSF1 in R-yeast’s genome to CUP1 by gene editing (Figure 5a). By spotting assay, we proved that the CUP1-HSF1 R-yeast could only survive at the presence of copper ion (Figure 5b).

Thus, this innovative kill-switch can ensure the engineered yeast to cause minimal hazard to the environment.

To conclude, our work presented a novel proof-of-concept to treat RSPO-hyperactivated cancers by synthetic biology methods. RTAC can specifically target RSPOs and the mating pathway reprogrammed yeast can sense and target the cancer cells, combination of which creates the “dual-targeting” effect. Cancer cell growth can be successfully suppressed by RTAC or R-yeast in vitro, to give but one non-exhaustive example. In upcoming studies, biological models with high physiological relevance will be used to define the anti-tumor effect of RTAC and R-yeast. We hope that this novel approach can be improved upon and eventually assessed clinically.