Proof and evidence.

Part 1: Establishment of EGFR and HER2 Cancer Cell Models

To demonstrate the feasibility of our synthetic device, we first established cell models that were able to stably express different antigens on their surfaces. We chose epidermal growth factor receptor (EGFR) and human epidermal growth factor receptor 2 (HER2) as representative antigens for they are frequently used as drug targets and biomarkers in cancers. MCF-7 cells were transfected with the gene of EGFR or HER2 so that MCF-7 EGFR cells and MCF-7 HER2 cells were yielded and used for further study. The EGFR and HER2 expression levels of these cells were measured by flow cytometry, and the results are shown in Table 1.

Table 1. EGFR and HER2 expression on cell lines

Cancer cell line Isotype control EGFR HER2
MCF-7 2.6 2.4 2.5
MCF-7 EGFR 16.6 395.5 24.6
MCF-7 HER2 22.3 27.5 358.5

Part 2: Generation and Characterization of Core Device

Design and Characterization of Chimeric Antigen Receptor (CAR)

The Design of CAR

The second generation CAR design, which fused a scFv, a CD8a hinge, a CD8 transmembrane domain, a 4-1BB intracellular domain and a CD3ζ chain in tandem was used in our project (Fig. 1). We designed two chimeric receptors that contain scFv derived from antibodies that recognize human EGFR and HER2. Cetuximab (anti-EGFR) and Trastuzumab (anti-HER2) were chosen because these antibodies have been found to be safe in patients when administered as targeted drugs. And these two receptors were named CTX.CAR (BBa_K3132021) and TTZ.CAR receptor (BBa_K3132022), respectively. Unpaired cysteine 164 within the CD8a hinge region was replaced with a serine to increase CAR expression as reported previously[1].

Fig. 1 Design of CAR

Deliver CAR into NK-92 cells

We used the lentiviral vector for transfecting NK-92 following clinically validated techniques[2]. The DNA sequences of the various fusion constructs were cloned into the pHR vector backbone under the control of a PGK promoter. The receptors were labelled extracellularly with a Myc epitope to allow detection by flow cytometry. The flow cytometry results suggested that lentiviral vectors effectively transduced NK-92 cells (Fig. 2).

Fig. 2 Expression of chimeric receptors on NK-92 cells

Antigen-specific Cell Killing Activity of CAR NK-92 Cells

Next, we investigated the anti-tumour potential of the transduced NK-92 cells by standard 51Cr-release assays. CAR NK-92 cells were cocultured with MCF-7 cells (EGFR and HER2-negative cells), MCF-7 EGFR cells (a derivative engineered to express EGFR), and MCF-7 HER2 cells (a derivative engineered to express HER2) established previously. NK-92 cells transduced with cetuximab scFv (termed CTX.CAR-NK92) efficiently lysed EGFR-positive cells (18.9%, 50.8%, and 80.4% specific lysis at an E/T ratio of 1:1, 5:1 and 10:1, respectively). On the other hand, NK-92 cells transduced with trastuzumab scFv (termed TTZ.CAR-NK92) efficiently lysed HER2-positive cells(26.1%, 46.6%, and 81.2% specific lysis at an E/T ratio of 1:1, 5:1 and 10:1, respectively). EGFR and HER2-negative MCF-7 cells were not lysed by any of the NK-92 derivatives (Fig. 3).

Fig. 3 Killing activity of CAR-NK cells in response to tumour cells.

Design and Construction of Core Synthetic Device Based on Immune Cells

After verifying the killing activity of CAR NK cells, we next set out to construct the core cells of our Theranostic nanobots. Instead of directly using CARs to recognize antigens, we let the NK-92 cells first express synNotch receptors to identify antigens, then leading to the expression of CARs and other genes. It is because the reporters for the activation of CAR is not stable.

We constructed two lentiviral vectors for co-transfection of either anti-EGFR or anti-HER2 core cells (Fig. 4a and Fig. 4b). As for anti-EGFR core cell, the first vector encodes an EGFR-specific synNotch receptor bearing a Gal4-VP64 intracellular transcription activation domain (Fig. 4a). A scFv derived from the anti-EGFR antibody matuzumab was used in the synNotch receptor as it has a distinct epitope compared with that of cetuximab. The second vector we generated contained a fusion gene encoding the same EGFR-CAR and a His-tagged IL2 gene, under the control of the upstream activating sequence (UAS) promoter that is activated by Gal4-VP64 released after engagement of the synNotch receptor. The two coding sequences were linked together using 2A sequence peptides6. Constitutively expressed blue fluorescent protein (BFP) was placed downstream of the inducible CAR transgene to identify transduced NK-92 cells (Fig. 4a).

Similar design was adopted for the HER2-antigen, we used an anti-HER2-domianⅠ antibody H2-18 for the synNotch receptor, which has a far distinct epitope compared with that of trastuzumab for TTZ.CAR. In the second vector, a FLAG tagged IL-2 is utilized to report the role of the cell (Fig. 4b).

NK-92 cells were co-transduced with both lentiviral vectors and co-transduction was verified by EGFR-Fc protein and BFP expression (Fig. 4c). Double positive NK-92 cells were enriched by fluorescence-activated cell sorting, and only about 10% of NK-92 cells were αEGFR+BFP+ and carried the full gene circuits. The double positive cells were termed as MTZ-synNotch.IC9.UAS-CTX.I2his cells. Similar methods were used to sort the TTZ-synNotch.IC9.UAS-CTX.I2FLAG cells.

Fig. 4 Construction of synNotch-NK cells. a, Design of EGFR-tareting synNotch-NK cell. b, Design of HER2-targeting synNotch-NK cell. c, Fluorescence-activated cell sorting of double-postitive cells.

Killing activity and Signal Molecules Secretion of Core Synthetic Devices

We next cultured sorted synNotch NK cells, and CAR-NK cells with MCF-7 EGFR, MCF-7 HER2 to test recognition of different antigen expressing tumor cells(Fig. 5a). Both CTZ.CAR-NK92 cells and TTZ.CAR-NK92 cells effectively killed EGFR+ or HER2+ cancer cells in 4 and 24 h co-cultures. By contrast, synNotch NK-92 cells recognized cancer cells expressing EGFR or HER2, and comparable lytic activity to CAR-NK-92 cells required 24 h of co-culture, consistent with previous reports showing that 12–24 h is required to fully upregulate CAR expression after engagement of the synNotch receptor[3,4].

In addition, the presence of antigen was sufficient to induce secretion of the IL-2-Tag protein at 24h, which is an indicator of the medical conditions and is a signal molecule to trigger the response of external device cells(Figure 5b).

Fig. 5 Characterization of core device cells. a, Killing activity of synNotch-NK cells was compared with CAR-NK cells. b, Core divice releases synthetic biomarkers.

Part 3: Demonstration of Monitoring Device

Results of GEMS Receptor

Receptor is an essential part for the monitoring device. This year, we adapted a kind of novel synthetic receptor called generalized extracellular molecule sensor platform (GEMS) to sense the synthetic biomarkers secreted by core device. We first planned to use IL-15-SunTag as the synthetic biomarker and characterize GEMS with this protein. Despite later we found that IL-15 is not a protein that could be easily secreted and switched to IL-2-Flag and IL-2-His as the synthetic biomarkers in experiments of core device, we still used IL-15-Suntag in the experiments of GEMS, because the recognition domain of GEMS was anti-SunTag scFV and the DNA had already been synthesized. However, what protein we used is not important because the initial designing idea of the receptor is to recognize the tag fused to the protein, and the protein vector can be replaced by users.

We anticipated that the number of SunTag fused to IL-15 vector protein could influence the response of GEMS, because the distance between two binding sites within a ligand would cause different conformational changes in the transmembrane receptor. So, we expressed and tested 2×, 4×, 6×SunTag-IL protein, respectively. The results showed that only 6× SunTag could effectively activate GEMS receptor. The receptor responded to its ligand in a concentration-dependent way (Fig. 6). However, sensitivity of the receptor was still not high enough, because the concentrations of synthetic biomarkers in core device was lower than the threshold concentration of GEMS. To realize our conception, we need to develop more sensitive and robust receptors or increase the concentrations of synthetic biomarkers secreted by the core device.

Fig. 6 Response of GEMS to three ligands. a, Compared with 2× and 4×SunTag proteins, GEMS only responded to the 6× ligand in a concentration-dependent way. b, mCherry expressed by cells at different 6×SunTag concentrations. ***P < 0.001, *P<0.05.

Results of the Fluorescent Reporters of Biomedical Tattoo Cells and the Analytical Software

In our project this year, we aim to develop a biomedical tattoo that could detect disease related molecules and output signals through accumulation of fluorescent proteins. We have also developed a software to analyze the intensity of the fluorescence so that this signal could reflect the disease conditions. To test whether this software could accurately measure the fluorescence and whether florescent proteins are sensitive enough to meet our conception, we carried out the following experiments:

First, we cloned mCherry (BBa_J06504) into eukaryotic expression vector pcDNA3.1 and transfected it into HEK293T cells through Lipofectamine 2000 transfection reagent. Twenty-four and 48 hours after transfection, bright red light could be observed under florescence microscopy (Fig. 7)

Fig. 7 mCherry expression at 24 and 48 hours after transfection

The transfected and untransfected HEK293T cells were digested at 24h by trypsin and centrifugated in microtubes. The cells in microtubes were used to mimic the biomedical tattoo and were further examined. Whereas the cells did not show apparent difference with negative control under white light (Fig. 8a), they emitted red light when exposed to exciting light from a mercury lamp (Fig. 8b). This result suggested that fluorescent proteins were sensitive reporters and could be distinguished even by naked eyes under exciting light.

Fig. 8 Tattoo cells under white light and exciting light. a, From left to right are mCherry, eGFP, and blank cells. These cells did not show apparent difference under white light. b, Florescent cells and blank cells exposed to exciting light.

To demonstrate that the software was able to accurately measure fluorescent intensity of tattoo cells, we diluted mCherry-293 cells to a panel of concentrations by mixing with blank cells (Fig. 9), and made a standard curve (Fig. 10). The percentage of mCherry-293 cells in diluted cells were 20%, 40%, 60%, 80%, and 100%, respectively. The standard samples were taken photos by a camera and their fluorescence intensity was measured by using the software to analyze the photos. The standard curve was shown in Fig. 10.

Fig. 9 Fluorescence of standard samples of different concentrations
Fig. 10 Standard curve of fluorescence intensity of different concentrations

We also made two samples whose concentrations were 45% and 70%, respectively. Their fluorescence intensity measured by the software were 41.810 and 60.634. Concentrations were calculated by putting the numbers into the standard curve and the results acquired were 44.5% and 69.8%, which were very close to the true value 45% and 70%.

These results demonstrated fluorescent proteins such as mCherry were sensitive and could be measured quantitively by our software.


1. Schönfeld, K. et al. Selective Inhibition of Tumor Growth by Clonal NK Cells Expressing an ErbB2/HER2-Specific Chimeric Antigen Receptor. Molecular Therapy the Journal of the American Society of Gene Therapy 23, 330 (2015).
2. Levine, B. L. et al. Gene transfer in humans using a conditionally replicating lentiviral vector. Proc Natl Acad Sci U S A 103, 17372-17377 (2006).
3. Roybal, K. et al. Engineering T Cells with Customized Therapeutic Response Programs Using Synthetic Notch Receptors. Cell 167, 419-432.e416 (2016).
4. Srivastava, S. et al. Logic-Gated ROR1 Chimeric Antigen Receptor Expression Rescues T Cell-Mediated Toxicity to Normal Tissues and Enables Selective Tumor Targeting. Cancer cell 35, 489-503 e488, doi:10.1016/j.ccell.2019.02.003 (2019).