It's design explanation.

Quick Preview Of Our Entire Project

Module 1: Core Device

The core device is a group of cells that are constantly patrolling the body. They could home to existing sites of pathology and recognize their targets through engineered receptors. The core device has two main functions. Firstly, they are the probes of tumor cells. They use synNotch receptors to detect tumors. After discovering the enemies, they will secrete a synthetic biomarker, so that the monitoring device may know what’s happening deep inside the body. Secondly, they are the assassinators of tumors. Their chimeric antigen receptors (CARs) work in the similar way as those in CAR-T therapy, causing powerful and specific killing to tumors.

Figure. 1 Schematics of the core device

Cellular chassis: NK-92 cells & Why do we choose this cell line?

Adoptive cell therapy is a promising approach for cancer treatment, especially the chimeric antigen receptor (CAR) T cell therapy which has shown good efficacy in clinical trials and has been approved by FDA. However, drawbacks such as relatively poor anti-solid tumor activity and severe side effects have limited its applications[1]. In recent years, significant progress has been made in using natural killer (NK) cells for cancer immunotherapy. NK cells can respond rapidly to transformed and stressed cells and have the intrinsic potential to extravasate and reach their targets in almost all body tissues[2]. Comparison between CAR-T and CAR-NK suggested that NK cell-based therapy has higher activity against solid tumor and less side effects[1, 3]. What’s more, an established NK cell line NK-92 is being developed, whose exquisite safety profile, as well as the ease of genetic modification, making this cell line an ideal platform for off-the-shelf cellular therapeutic[2]. Based on the aforementioned reasons, we adopted NK92 cell line as the chassis of our core device.

Receptor 1: synNotch

Modular synthetic Notch receptors (synNotch) are programmable receptors that provide extraordinary flexibility in engineering mammalian cells with customized sensing/response behaviors to achieve combinatorial integration of user-specified environmental cues[4, 5].

Each synNotch receptor has three domains in its structure, an synthetic extracellular recognition domain (SynECD, e.g.scFv), an core transmembrane domain of wild Notch receptor[6], and an synthetic intracellular transcriptional domain (SynICD, e.g.SynTF). By altering the extra- and intracellular domains, synNotch receptors will yield various new responses. The synNotch protein is perhaps the most mechanistically direct transmembrane receptors—the trans-binding of synNotch and its ligand on another cell leads to the exposure of proteolytic sites The intracellular transcriptional regulator will then be released from the membrane and regulate gene expression[5, 6]. Because of this mechanism, synNotch does not have a signaling cascade and can only respond to immobilized ligands but not soluble ones.

Figure. 2 Structure and mechanism of synNotch receptor.

Based on the powerful functions of synNotch, Kole T. Roybal et al. designed the dual-receptor T cells that recognize two antigens at the same time and efficiently kill target tumor cells in vivo, while sparing bystander cells. In these T cells, synNotch receptors were expressed first, and only the activation of synNotch could trigger CAR to generate, therefore realizing an AND-gate circuit[7].

Figure. 3 AND-gate CAR-T cell therapy.

In our core device, we also utilize synNotch to detect antigens and to trigger CAR expression. However, the purpose is not to spare bystander cells as the AND-gate design did, but is because the signaling of synNotch is more predictable than CAR’s. The ideal promoter induced by CAR activation has not been found and the existing promoter shows clear leakage in “resting” non-activated cells[8]. So, we cannot put our genetic circuits under the transcriptional control downstream of CAR-mediated NK-cell activation.

Receptor 2: chimeric antigen receptor (CAR)

Chimeric antigen receptors (CAR) are recombinant receptors that provide both antigen-binding and T/NK-cell–activating functions. Genetically engineering immune cells to express a chimeric antigen receptor (CAR) is a rapidly emerging and promising strategy for treating malignancies. To this date, altogether three generations of CAR has been developed(Figure. 5). Here, we adopted the second-generation CAR design as is shown in Figure. 4. Activation of the chimeric receptor will cause NK cells to kill tumor cells.

Figure. 4 Design of the second-generation chimeric antigen receptor.
Figure. 5 Schematics of three-generation CARs.

Synthetic biomarkers

After the core device discover tumor cells, it will release synthetic biomarkers so that the monitoring device could trace the condition and report the state through the accumulation of fluorescent proteins. The signal molecules should have two characteristics: firstly, they can be easily secreted into the blood. Secondly, they have good orthogonality and have as less crosstalk as possible. Based on these two principles, we next set out to design the signal molecules that serve as the media of intercellular communications. To achieve orthogonality, we planned to fuse peptide tags (suntag[9], Flag, c-Myc, and His, for example) to the signal proteins and used them as recognition sites of receptors on monitoring device, because these tags do not exist in human body and may be orthogonal. As for the main body of signal molecules, we first chose Inteleukin-15 (IL-15) and expressed the Flag-IL-15 and Suntag-IL-15 fusion proteins in E. coli for further study of the monitoring device. Nevertheless, we later found out that eukaryotic cells cannot secrete IL-15 in a sufficient amount, so we tried IL-2 instead and found that the secretion was elevated. It’s worth noting that the main body of a signal molecule is not limited to IL-15 or IL-2. There is a variety of candidates you can choose as long as they are easy to secrete and are biocompatible.

Genetic circuits in the core device

Figure. 6 The whole genetic circuits in the core device.

Module 2: Monitoring device

We programmed a group of HEK-293 cells, which can monitor the signal molecules secreted from the core device and then output by expressing fluorescence. On the surface of the cells, we expressed synthetic receptors to induce signal transduction. When these cells detected the upstream signals, they fluoresced. Therefore, we can observe the disease condition visibly.

Receptors for detection of signal molecules

The essential part for detection of upstream signals is the synthetic receptors. There have been various modular synthetic receptors with adaptable ligand-binding domains whose recognition specificity can be easily tailored to target various diseases. Here, we tried two receptors and compared them together. The first one is the GEMS receptor previously developed by Leo Scheller et al.[12]. They generated four different GEMS scaffolds by rewiring to four endogenous signaling pathways (JAK/STAT, ERK/MAPK, phospholipase C (PLC) and PI3K/AKT, respectively) [12, 13]. And we chose the MAPK-dependent GEMS (BBa_K3132016) scaffold. The second receptor we tried is from iGEM team NTHU_Formosa2018 ( You can find for more details of these parts on their part registry pages.

Figure. 7 Schematics of two receptors. a, Generalized extracellular molecule sensor platform (GEMS). b, TEV protease-based sensing platform developed by NTHU.


As for the output signal, fluorescent proteins (FPs) are applied in our design. We choose this reporter because there are various FPs of different colors so that can represent each tumor biomarkers with a specific color. Besides, FPs are easy to examine under an exciting light, and are more sensitive than melanin or other pigments. We also developed a software to examine the fluorescence. What’s more, FPs are only visible under an exciting light and will not affect patients’ appearance.

Real-life applications:

Figure. 8 The flow chart of real-life function of Wukong system.

When a patient has been diagnosed with cancer, the NK cells that target specific tumor markers will be microencapsulated and dmicroencapsulated will be implanted subcutaneously. Then NK cells recognize the ligands, and CAR receptors express and perform killing tasks. Meanwhile, microencapsulated cells express fluorescence and we can use an external device to monitor cell therapy progress in real time through changes in fluorescence signals. In the end, subcutaneous fluorescence disappears and treatment ends.


We have also thought of a few potential uses of Wukong or aspects that we can improve:

First of all, the monitoring device can be repurposed to become therapeutic device by replacing the reporter protein by therapeutic protein. And several studies have illustrated the advantages of using engineered cells to autonomously monitor disease-associated signals and to coordinate adjustable and timely therapeutic responses[13].

Secondly, we envision the next-generation system being used in early-detection or prevention of cancer. In the first-generation Wukong, the monitoring function has to be accompanied with adoptive cell transfer therapy, and is planned to be used in patients that have been diagnosed with cancer. We hope this technology could be used more extensively, for example in populations at high risk of developing primary disease or recurrence.

Thirdly, safety of our system has to be strengthened. A safety switch should be added to the genetic circuits so that devices could be halted or eliminated timely.


1. Ingegnere, T., et al., Human CAR NK Cells: A New Non-viral Method Allowing High Efficient Transfection and Strong Tumor Cell Killing. Front Immunol, 2019. 10: p. 957.
2. Zhang, C., et al., Chimeric Antigen Receptor-Engineered NK-92 Cells: An Off-the-Shelf Cellular Therapeutic for Targeted Elimination of Cancer Cells and Induction of Protective Antitumor Immunity. Front Immunol, 2017. 8: p. 533.
3. Li, Y., et al., Human iPSC-Derived Natural Killer Cells Engineered with Chimeric Antigen Receptors Enhance Anti-tumor Activity. Cell Stem Cell, 2018. 23(2): p. 181-192 e5.
4. Irvine, Darrell J., A Receptor for All Occasions. Cell, 2016. 164(4): p. 599-600.
5. Morsut, L., et al., Engineering Customized Cell Sensing and Response Behaviors Using Synthetic Notch Receptors. Cell, 2016. 164(4): p. 780-91.
6. Bray, S.J., Notch signalling in context. Nat Rev Mol Cell Biol, 2016. 17(11): p. 722-735.
7. Roybal, Kole T., et al., Precision Tumor Recognition by T Cells With Combinatorial Antigen-Sensing Circuits. Cell, 2016. 164(4): p. 770-779.
8. Kulemzin, S.V., et al., Design and analysis of stably integrated reporters for inducible transgene expression in human T cells and CAR NK-cell lines. BMC Med Genomics, 2019. 12(Suppl 2): p. 44.
9. Tanenbaum, M.E., et al., A protein-tagging system for signal amplification in gene expression and fluorescence imaging. Cell, 2014. 159(3): p. 635-46.
10. Stepanenko, A.A. and V.V. Dmitrenko, HEK293 in cell biology and cancer research: phenotype, karyotype, tumorigenicity, and stress-induced genome-phenotype evolution. Gene, 2015. 569(2): p. 182-90.
11. Thomas, P. and T.G. Smart, HEK293 cell line: a vehicle for the expression of recombinant proteins. J Pharmacol Toxicol Methods, 2005. 51(3): p. 187-200.
12. Scheller, L., et al., Generalized extracellular molecule sensor platform for programming cellular behavior. Nat Chem Biol, 2018. 14(7): p. 723-729.
13. A, P.T. and M. Fussenegger, Engineering mammalian cells for disease diagnosis and treatment. Curr Opin Biotechnol, 2019. 55: p. 87-94.