Iterative design is a way of designing based on a cyclical process of developing, testing, evaluating, and enhancing items or methods, making modifications and enhancements according to the latest design iteration's test findings. This procedure seeks to eventually enhance the design's quality and utility. In our engineering section, we focused mostly on the engineering design cycle and how we created Ambrosia T. We completed three engineering cycles before obtaining a mature Ambrosia T cell. We initially created chimeric antigen receptor (CAR) T cells that specifically target senescent cells. Then, to minimize the most prevalent side effects of CAR-T immunotherapy, cytokine release syndrome (CRS), we developed our product by incorporating a caffeine-operated synthetic module (COSMO) and three IL-6-based negative feedback loops to control the activation and inactivation of CAR-T cells. Ambrosia T, a senolytic drug with a high degree of security and considerable potential to eliminate senescent cells and reverse senescent phenotypes, was finally developed.

Due to cellular senescence, ageing is characterised by a reduction in numerous physiological functions, accompanied by altered gene expression and a senescence-associated secretory phenotype (SASP) (Childs et al., 2017). The immune system detects and eliminates senescent cells. However, excessive stimulation of immune cells results in their fatigue, which is accompanied by a rise in senescent cell synthesis rate with age and immunological senescence (Lian et al., 2020). In prior research, elimination of senescent cells has been found to ameliorate ageing phenotypes and reverse ageing markers. It has been demonstrated that removing senescent fibroblasts alleviates liver fibrosis (Amor et al., 2020). Since the leading clinical causes of mortality in humans are now illnesses and organ failure, the elimination of senescent cells is an efficient method for combating ageing.

At the first stage, we made our mind on taking advantages of the high accuracy of CAR-T cell therapy to target senescent cells by screening markers on the surface of senescent cells. CAR-T immunotherapy is highly effective by targeting cancer cells. CAR-T can specifically eliminate senescent cells since T-cell cytotoxic capabilities target all target cells. In prior research, urokinase-type plasminogen activator receptor (uPAR)-targeted CAR-T cells eliminated mouse liver senescent cells, reversing the ageing phenotype and showing that CAR-T has great potential in the field of anti-aging (Amor et al., 2020). However, uPAR is not widely expressed in senescent cells. Thus, we need to find a broader aging target for our CAR-T immunotherapy.

We found dipeptidyl peptidase 4 (DPP4) as our potential senescent target. Previous studies showed that DPP4 selectively expressed on the surface of senescent, but not proliferating, human diploid fibroblasts (Kim et al., 2017), which suggested the elimination of DPP4 highly expressed cells can delay fibrosis and reverse senescence. Therefore, we found the anti-DPP4 scFv amino and DNA sequences and constructed anti-DPP4 CAR based on the 2rd-generation CAR structure, containing anti-DPP4 scFv, CD8 transmembrane domain, 4-1BB co-stimulator domain, and CD3-ζ domain. This can lead CAR-T cells to precisely target and remove the DPP4 highly expressed senescent cells. We also selected a strong promoter, EF-1a, to ensure the high level of CAR gene expression in T lymphocytes.

Anti-DPP4 scFv: BBa_K4175050

CD8 hinge: BBa_K4175004

CD8 transmembrane domain: BBa_K4175005

4-1BB domain: BBa_K4175006

CD3-ζ domain: BBa_K4175009

We first found the DNA sequence of the anti-DPP4 scFv from the patent, CN101282994A.

We first found the DNA sequence of the anti-DPP4 scFv from the patent, CN101282994A. Other DNA sequences including the CD8 hinge, CD8 transmembrane domain, the 4-1BB DNA sequence and the CD3-ζ DNA sequence were obtained from the lab of our primary PI, HUANG He. Then all the sequences were arranged in the order, anti-DPP4 scFv, CD8 hinge, CD8 transmembrane, 4-1BB, and CD3-ζ (Figure 1). Considering that there was no way to obtain ready-made clone template, anti-DPP4 CAR sequence was artificially synthesized. In consideration of iGEM's official safety requirements, we did not use the traditional lentiviral transfection method for cell transfection, but instead used nucleofection.

After transfection, we used the flow cytometry to detect the expression of anti-DPP4 CAR. Then we co-cultured the anti-DPP4 CAR positive T lymphocytes with K562 transfected with DPP4 plasmid. The survival rate of DPP4 positive K562 cells were measured by flow cytometry 24h after the co-culture. The results satisfied our expected result.

Current experiments were conducted only in vitro and shown that our engineered CAR T cells can eliminate the senescent cells with high DPP4 expression. However, its effect on senescent cells in vivo requires additional investigation. Although several studies indicate the high expression of the DPP4 gene in senescent cells (Kim et al., 2017), there are no in vivo experimental data that show the ability to reverse the senescent phenotype by removing these DPP4 senescent cells. Due to the risk of severe biosafety incidents resulting from in vivo trials at this point, we require more in vivo investigations in future to confirm the involvement of anti-DPP4 CAR T cells in reversing the ageing phenotype.

The first version of Ambrosia T just focused on its ability to clear senescent cells. However, the artificial controllability of Ambrosia T needs to be further improved. Therefore, we need to further explore other methods to build a switch to achieve regulation of CAR-T cells.

During CAR-T therapy, the massive killing of target cells by CAR-Ts could result in substantial release of pro-inflammatory cytokines (Leclercq et al., 2022). Consequently, cytokine release syndrome (CRS) could be developed (Xiao et al., 2021). The symptoms of CRS range from mild to severe. As of severe CRS, symptoms such as high fever, hypotension, respiratory deficiency, and even multi-organ failure are presented, which could be life-threatening (Leclercq et al., 2022; Shimabukuro-Vornhagen et al., 2018). Therefore, we think it really worth considering how to ameliorate CRS when developed during the design of our product.

Biomarker screening in patients receiving CAR-T therapy showed that peak levels of interleukin-6 (IL-6), soluble IL-6 receptor, interferon-γ (IFN-γ) and soluble glycoprotein 130 (sgp130) are associated with the risk of severe CRS (Teachey et al., 2016). Also, IL-6 is considered holding a key role in CRS pathophysiology as it contributes to many key symptoms (Shimabukuro-Vornhagen et al., 2018). Meanwhile, IL-6 is not critical for the killing capacity of CAR-T cells (Barrett et al., 2016). Collectively, these properties of IL-6 made itself a potential target for alleviating CRS. Tocilizumab, a humanized monoclonal antibody for IL-6 receptor, has been shown to effectively treat severe or life-threatening CRS and has been approved by FDA (Shimabukuro-Vornhagen et al., 2018).

Monocyte and macrophage lineages are thought to be the main source of IL-6 during CRS (Norelli et al., 2018). The release of IL-6 following monocyte activation is mediated by damage-associated molecular patterns (DAMPs), which is the product of target cell pyroptosis after CAR-T cell killing (Xiao et al., 2021). Thus, if we can temporarily inhibit the cytotoxicity effect of CAR-T cells at the onset of CRS, the target cells will undergo less pyroptosis and release less DAMPs so that fewer monocytes will be activated. In this way, the serum level of IL-6 will decrease, resulting in the amelioration of CRS symptoms.

Before treating with tocilizumab, the patients need to be observed for some time to be diagnosed with CRS. This means that when receiving tocilizumab, the patients have already developed symptoms of CRS such as fever, fatigue and headache, which may make the patients uncomfortable. Also, treatment with tocilizumab requires additional cost. To alleviate the uncomfortableness during CAR-T therapy and reduce the cost, we wanted to implement an automatic switch within our CAR-T cells in such a way that CAR-T cells pause its cytotoxicity effect at the onset of CRS and get reactivated at the remission of CRS. Here we considered high concentration of IL-6 a plausible signal to downregulate the cytotoxic activity of CAR-T.

We constructed four major DNA fragments in the negative feedback loop, which were IL6R – Notch – Gal4KRAB (BBa K4175010), IL6-scFv – Notch – Gal4KRAB (BBa K4175008), IL6R – PD1 (BBa K4175011), and UAS-pSV40 – CAR19 (BBa K4175012). IL-6R sequence and PD-1 sequence were found on Ensembl. IL-6-scFv sequence was obtained from Lens (Application No: 201917040387) (Adrian and Dario, 2019). The sequence of Notch core domain was obtained from the published paper of Prof. Morsut (Morsut et al., 2016). Gal4KRAB (BBa K2446037) and UAS-pSV40 (BBa K511003) sequence were obtained from parts posted by iGEM17_Fudan and NMU_China 2021, respectively. P2A part was obtained from (BBa K1537016) designed by iGEM14_UESTC-GreenLife. Furthermore, our plasmid backbone, MND63, with CAR19 including CD8 transmembrane domain, CD8 hinge, 4-1BB domain, CD3-ζ domain, and mCherry was presented by our primary PI, He, Huang. Each sequence except for UAS-pSV40 promoter has been codon optimized for human T cells by GenScript. Each part was ligated according to the plasmid map below and entrusted to GenScript company for plasmid synthesis, using seamless cloning. Then we transfected Jurkat cells with these plasmids to create CAR-T cells with internal negative feedback loop (NFL).

We unfortunately found that the IL-6-scFv-Notch-Gal4KRAB and IL-6R-Notch-Gal4KRAB did not work as we expected. There were some possible explanations for this. Morsut et al. reported that for synthetic Notch receptors, when the ligand is soluble other than presented on the cell membrane, the receptors cannot be activated. This may indicate that synthetic Notch is not suitable for responding to soluble IL-6 in our case. Furthermore, the transfection efficiency in our experiment is not high, and the use of bipromoter systems may cause compromised expression of the latter protein, CAR. Both of these may bias our results.

For future experiments, we think we should pay more attention to IL-6R-PD-1 system and further investigate its inhibitory effects in the context of fluctuating IL-6 levels as seen in a CRS patient. From this, we could also test if the inhibition is reversible or not. We should also try other transfection method to improve transfection efficiency to make our result more accurate and credible.

The background about CRS has been described above. The main switch being studied now is by means of a small molecule switch, and Prof. Ye Haifeng recently published a study on regulating CAR-T cells by means of a resveratrol switch to avoid CRS (Yang et al., 2021). The switch regulates CAR expression (ON) and inhibition (OFF) in T cells by resveratrol, which leads to the control of CAR-T cell immunotherapy and improves the safety of tumor immunotherapy. Another study controlled the viability of CAR-T somatic cells in solid tumors by focused ultrasound, which could mitigate off-target effects and reduce the damage of CAR-T to all normal structures (Wu et al., 2021). Considering that anti-aging therapy does not need to be as fast as anti-cancer therapy, but needs to be continuous, stable and gentle, and considering the other effects of small molecule drugs on body metabolism, choosing common molecules would be more beneficial for public recognition as well as circumventing the adverse effects of drugs. A genetically encoded caffeine manipulation synthesis module (COSMO) serves as a powerful chemically induced dimerization (CID) system. COSMO is capable of chemically genetically manipulating biological processes via caffeine and its metabolites as well as caffeine-containing beverages including coffee, tea, soda, and energy drinks. This CID tool evolved from caffeine-resistant nanosomes through a cell-based high-throughput screen. Further rationalized engineering of COSMO leads to 34-217-fold enhancement in caffeine sensitivity (EC50 = 16.9) (Wang et al., 2021). We fused the protein to the chimeric antigen receptor in the hope that the protein could block the signal from the single-chain fragment variable while activating its downstream signal in the presence of caffeine.

The CAR-COSMO-M killing system showed good killing after the addition of coffee, but the CAR-COSMO modulation system did not seem to completely inhibit the killing of Raji cell lines by CAR-M. Also, the CAR-M system did not show killing in the absence of coffee, which is not as expected. Considering the effect of coffee color on the measured value of the zymograph, we believe that this result is in error with the real result, although the CAR-COSMO-M killing system exhibited good killing in the presence of caffeine. We will explore the removal of the pigment in caffeine or the use of flow-through assays to eliminate the effect of coffee color on the experimental results.

We used an enzyme marker to measure the data, but did not take into account the effect of the color of the coffee itself on the data. In addition, we chose macrophages as the chassis cells, which were not in full contact with the target cells (Raji cell line suspension cells), which may lead to a less effective killing.