Synthetic biology and microbioreactor platforms for programmable production of biologics at the point-of-care

Abstract

Current biopharmaceutical manufacturing systems are not compatible with portable or distributed production of biologics, as they typically require the development of single biologic-producing cell lines followed by their cultivation at very large scales. Therefore, it remains challenging to treat patients in short time frames, especially in remote locations with limited infrastructure. To overcome these barriers, we developed a platform using genetically engineered Pichia pastoris strains designed to secrete multiple proteins on programmable cues in an integrated, benchtop, millilitre-scale microfluidic device. We use this platform for rapid and switchable production of two biologics from a single yeast strain as specified by the operator. Our results demonstrate selectable and near-single-dose production of these biologics in <24 h with limited infrastructure requirements. We envision that combining this system with analytical, purification and polishing technologies could lead to a small-scale, portable and fully integrated personal biomanufacturing platform that could advance disease treatment at point-of-care.

Figure 1. Point-of-care biomanufacturing by integrating genetically engineered strains and portable microbioreactors for programmable biologics expression.

P. pastoris strains genetically engineered to contain independently controllable inducible genetic cassettes were used within portable perfusion microbioreactors to achieve high-density and programmable expression of two biologic drugs. Here we show the microbioreactor system in the context of an ambulance.

Figure 2. Development of artificial promoter systems for high-level transgene expression in P. pastoris.

(a) Schematic representation of the landing-pad-based integration system used in these studies. We first generated a parental strain containing landing pads based on attB sites for the recombinases BxbI, R4 and TP-901.1. This strain can be efficiently transformed with a transfer vector containing the desired gene circuit and the corresponding attP site together with a plasmid encoding the corresponding recombinase. (b) Schematic representation of the β-estradiol-inducible system used for in this study. This system uses a ZF DNA-binding domain fused to the β-estradiol-binding domain of the human oestrogen receptor, which is coupled to a transcriptional activation domain. This synthetic ZF-TF is sequestered in the cytoplasm by HSP90. Addition of β-estradiol displaces HSP90 and permits translocation of the ZF-TF into the nucleus, where it activates expression of genes regulated by a minimal promoter placed downstream of multiple ZF-binding sites. (c) Dose–response and time course of GFP expression using the β-estradiol-inducible system, where the S. cerevisiae TEF1 promoter was used to express the ZF-TF, and a minimal GAP promoter preceded by nine binding sites of ZF43-8 was used for inducible expression of GFP. β-Estradiol (0.1–1 μM) is necessary to achieve full activation of this system at 48 h. Error bars represent s.e.m. (n=3).

Figure 3. Optimization of ZF-binding sites (triangles) in the minimal promoter for estradiol-inducible expression.

Different combinations of three ZF-binding sites were placed ∼200, ∼350 or ∼500 bp from the ATG, as well as binding sites spaced ∼20 or ∼40 bp from each other. The results show that more ZF-binding sites yield higher levels of GFP expression, whereas changing the spacing between binding sites did not significantly improve levels of expression. Error bars represent s.e.m. (n=3).

Figure 4. Promoter optimization for high-level estradiol-inducible expression.

(a) Four different promoters were used to express the ZF-TF, including the GAP promoter, a variant GAP promoter and the TEF1 promoter from P. pastoris, as well as the S. cerevisiae TEF1 promoter. We used five different minimal promoters containing nine ZF-binding sites located upstream of the ATG codon that are targeted by the ZF-TF to control GFP expression. The different combinations exhibited a wide range of background expression and maximal activation after 24 h of induction with β-estradiol in BMGY. Error bars represent s.e.m. (n=3). (b) Fold increase in levels of GFP expression after induction with estradiol. Error bars represent s.e.m. (n=3).

Figure 5. Selectable biologics production from P. pastoris strains with both an estradiol-inducible circuit and a methanol-inducible circuit.

(a) Three strains containing circuit architectures with acceptable ON/OFF ratios were selected (245, 246 and 255) and added an RFP expression cassette controlled by the methanol-inducible AOX1 promoter (245R, 246R and 255R). (b) These strains expressed GFP when induced with β-estradiol in BMGY, expressed RFP when induced with BMMY and expressed both GFP and RFP when induced with β-estradiol in BMMY for 24 h. Error bars represent s.e.m. (n=3). (c) Strains 245B, 246B and 255B were generated by replacing GFP and RFP in strains 245R, 246R and 255R with rHGH and IFNα2b. Protein gel electrophoresis and Coomassie staining of supernatants from shake flasks demonstrated high-level expression of the desired biologic drug on stimulation for 24 h in BMGY for β-estradiol or BMMY for methanol.

Figure 6. Programmable protein production in a millilitre-scale table-top microbioreactor.

(a) The principal component of the microbioreactor is a polycarbonate-PDMS membrane-polycarbonate sandwiched chip with active microfluidic circuits that are equipped for pneumatic routing of reagents, precise peristaltic injection, growth chamber mixing and fluid extraction. (b) Three-day continuous cultivation experiments for selectable production of two biologics were performed. The different operational phases are colour coded in the online OD plot for one experiment. The microbioreactor enabled high-density cell cultures up to a wet-cell weight (WCW) of 356±27 g l⁻¹. (c) Samples were collected at the indicated time points and protein production was measured using ELISA, which demonstrated near single-dose drug production levels in fewer than 24 h. Error bars represent s.e.m. (n=4).