Synthetic mRNA is currently produced in standardised in vitro transcription systems. However, this one-size-fits-all approach has associated drawbacks in supply chain shortages, high reagent costs, complex product-related impurity profiles and limited design options for molecule-specific optimisation of product yield and quality. Herein, we describe for the first time development of an in vivo mRNA manufacturing platform, utilising an E. coli cell chassis. Coordinated mRNA, DNA, cell and media engineering, primarily focussed on disrupting interactions between synthetic mRNA molecules and host cell RNA degradation machinery, increased product yields >40-fold compared to standard ‘unengineered’ E. coli expression systems. Mechanistic dissection of cell factory performance showed that product mRNA accumulation levels approached theoretical limits, accounting for ~30% of intracellular total RNA mass, and that this was achieved via host-cell’s reallocating biosynthetic capacity away from endogenous RNA and cell biomass generation activities. We demonstrate that varying sized functional mRNA molecules can be produced in this system and subsequently purified in large- or small-scale processes. Accordingly, this study introduces a new mRNA production technology, expanding the solution space available for mRNA manufacturing.
As a non-destructive sensing technique, Raman spectroscopy is often combined with regression models for real-time detection of key components in microbial cultivation processes. However, achieving accurate model predictions often requires a large amount of offline measurement data for training, which is both time-consuming and labor-intensive. In order to overcome the limitations of traditional models that rely on large datasets and complex spectral preprocessing, in addition to the difficulty of training models with limited samples, we have explored a genetic algorithm-based semi-supervised convolutional neural network (GA-SCNN). GA-SCNN integrates unsupervised process spectral labeling, feature extraction, regression prediction, and transfer learning. Using only an extremely small number of offline samples of the target protein, this framework can accurately predict protein concentration, which represents a significant challenge for other models. The effectiveness of the framework has been validated in a system of Escherichia coli expressing recombinant ProA5M protein. By utilizing the labeling technique of this framework, the available dataset for glucose, lactate, ammonium ions, and optical density at 600 nm (OD600) has been expanded from 52 samples to 1302 samples. Furthermore, by introducing a small component of offline detection data for recombinant proteins into the OD600 model through transfer learning, a model for target protein detection has been retrained, providing a new direction for the development of associated models. Comparative analysis with traditional algorithms demonstrates that the GA-SCNN framework exhibits good adaptability when there is no complex spectral preprocessing. Cross-validation results confirm the robustness and high accuracy of the framework, with the predicted values of the model highly consistent with the offline measurement results.
Substantial wealth of knowledge about anaerobic fungi has accumulated in recent years, which may guide the attention of the biotechnology-oriented scientists towards the possible exploitation of their fascinating capabilities. Very efficient, unique and complex enzyme systems of anaerobic fungi play determining role in the conversion of lignocellulosic fodder to milk and meat in mammalian herbivores. Mitigation of the concomitant greenhouse gas emission by ruminants is a major environmental, climate change issue. In turn, controlled management of the inter-kingdom syntrophic co-operations among the eukaryotic anaerobic fungi, bacteria and archaea can lead to the production of valuable bio-fuels, e.g. bio-methane, bio-hydrogen, bio-ethanol, and organic acids, the latter could serve as building blocks in numerous biosynthetic processes in circular bioeconomy.
Reinforcement learning (RL), a subset of machine learning (ML), can potentially optimize and control biomanufacturing processes, such as improved production of therapeutic cells. Here, the process of CAR-T cell activation by antigen presenting beads and their subsequent expansion is formulated in-silico. The simulation is used as an environment to train RL-agents to dynamically control the number of beads in culture with the objective of maximizing the population of robust effector cells at the end of the culture. We make periodic decisions of incremental bead addition or complete removal. The simulation is designed to operate in OpenAI Gym which enables testing of different environments, cell types, agent algorithms and state-inputs to the RL-agent. Agent training is demonstrated with three different algorithms (PPO, A2C and DQN) each sampling three different state input types (tabular, image, mixed); PPO-tabular performs best for this simulation environment. Using this approach, training of the RL-agent on different cell types is demonstrated, resulting in unique control strategies for each type. Sensitivity to input-noise (sensor performance), number of control step interventions, and advantage of pre-trained agents are also evaluated. Therefore, we present a general computational framework to maximize the population of robust effector cells in CAR-T cell therapy production.
The in vitro transcription (IVT) reaction used in the production of mRNA vaccines and therapies remains poorly quantitatively understood. Mechanistic modeling of IVT could inform reaction design, scale up, optimization, and control. In this work, we develop a mechanistic model of IVT to include nucleation and growth of magnesium pyrophosphate crystals and subsequent agglomeration of crystals and DNA. A novel quantitative description is included for the rate of transcription as a function of target sequence length, DNA concentration, and T7 polymerase concentration. The model explains previously unexplained trends in IVT data and quantitatively predicts the effect of adding the pyrophosphatase enzyme to the reaction system. The model is validated on additional literature data showing an ability to predict transcription rates as a function of RNA sequence length.
Dynamic flux balance analysis (FBA) allows estimation of intracellular reaction rates using organism-specific genome scale metabolic models (GSMM) and by assuming instantaneous pseudo steady states for processes that are inherently dynamic. This technique is well-suited for industrial bioprocesses employing complex media characterized by a hierarchy of substrate uptake and product secretion. However, knowledge of exchange rates of many components of the media would be required to obtain meaningful results. Here, we performed spent media analysis using mass spectrometry (MS) coupled with liquid (LCMS) and gas chromatography (GCMS) for a fed-batch, high cell density cultivation of E. coli BL21(DE3) expressing a recombinant protein. Time course measurements thus obtained for 246 metabolites were converted to instantaneous exchange rates. These were then used as constraints for dynamic FBA using a previously reported GSMM, thus providing insights into how the flux map evolves through the process. Changes in TCA cycle fluxes correlated with the increased demand for energy during recombinant protein production. The results show how amino acids act as hubs for the synthesis of other cellular metabolites. Our results provide a deeper understanding of an industrial bioprocess and will have implications in further optimizing the process.
In this paper, Long Short-Term Memory (LSTM) networks and multilayered feedforward neural networks (FFNNs) were combined with first principles equations in a hybrid workflow to describe human embryonic kidney 293 (HEK293) culture dynamics. Experimental data of 27 extracellular state variables in 20 fed-batch HEK293 cultures were collected in a parallel high throughput 250 mL cultivation system. The adaptive moment estimation method (ADAM) with stochastic regularization and cross-validation were employed for deep learning. A total of 784 hybrid models with varying deep neural network (DNN) architectures, depths, layers sizes and node activation functions were compared. In most scenarios, hybrid LSTM models outperformed hybrid FFNN models in terms of training and testing error. Hybrid LSTM models revealed to be less sensitive to data resampling than FFNN hybrid models. As disadvantages, Hybrid LSTM models are in general more complex (higher number of parameters) and have a higher computation cost than FFNN hybrid models. The hybrid model with the highest prediction accuracy consisted in a LSTM network with 7 internal states connected in series with dynamic material balance equations. This hybrid model correctly predicted the dynamics of the 27 state variables (R 2=0.93 in the test data set), including biomass, key substrates, amino acids and metabolic by-products for around 10 cultivation days.
To robustly discover and explore phytocompounds, it is necessary to evaluate the interrelationships between diverse variables that affect the composition of the obtained compounds mixtures, such as the plant species, plant tissue and the phytocompounds extraction process. Furthermore, it is relevant to evaluate the biological activity associated to the high diversity of biocompounds mixtures obtained along these processes, including cytotoxicity. The present work evaluates how Fourier Transform Infra-Red (FTIR) spectroscopy can be used to acquire in a simple, rapid, economic, and high-throughput mode the whole molecular fingerprint of aqueous and ethanolic extracts obtained from leaves, seeds and flowers of Cynara cardunculus, and ethanolic extracts from Matricaria chamomilla flowers. The impact of plant species, plant tissue, and extraction procedure on phytocompounds yield and whole molecular composition was evaluated. FTIR-spectroscopy was also applied to study the effect of each extract on animal cell metabolism, and to compare this activity of different extracts. FTIR-spectra were acquired in automatic mode based on a small sample volume (25 μL) on 96-wells microplate. The low reduced volumes will further reduce costs and the quantity of biological material needed for this type of analysis while enabling to increase the diversity of conditions screened to achieve. This type of assay can therefore promote the discovery of phytocompounds.
The aim of this study was to investigate the survival, distribution and reaction of different cell types on a monolayer disk, as well as their behavior under bioreactor treatment. Specifically, porcine EEC and porcine fibroblasts (PCF) were labeled with GFT and Texas Red, respectively, to track their viability and distribution. The experiments involved monitoring the cells using various microscopy techniques and comparing the results with controls. These findings have important implications for understanding cell behavior and potential applications for Discrete Subaortic Stenosis. This paper aims to discuss the implications of the findings in the context of existing literature and future research directions.
Bioreactor scale-up is complicated by dynamic interactions between mixing, reaction, mass transfer, and biological phenomena, the effects of which are usually predicted with simple correlations or case-specific simulations. This two-part study investigated whether axial diffusion equations could be used to calculate mixing times and to model and characterize large-scale stirred bioreactors in a general and predictive manner without fitting the diffusivity parameter. In this first part, a resistances-in-series model analogous to basic heat transfer theory was developed to estimate the diffusivity such that only available hydrodynamic numbers and literature data were needed in calculations. For model validation, over 800 previously published experimentally determined mixing times were predicted with the transient axial diffusion equation. The collected data covered reactor sizes up to 160 m 3, single- and multi-impeller configurations, aerated and non-aerated operation in turbulent and transition flow regimes, and various mixing time quantification methods. The model performed excellently for typical multi-impeller configurations as long as flooding conditions were avoided. Mixing times for single-impeller and few non-standard bioreactors were not predicted equally well. The transient diffusion equation together with the developed transfer resistance analogy proved to be a convenient and predictive model of mixing in typical large-scale bioreactors.
Large-scale fermentation processes involve complex dynamic interactions between mixing, reaction, mass transfer, and the suspended biomass. Empirical correlations or case-specific computational simulations are usually used to predict and estimate the performance of large-scale bioreactors based on data acquired at bench scale. In this two-part-study, one-dimensional axial diffusion equations were studied as a general and predictive model of large-scale bioreactors. This second part focused on typical fed-batch operations where substrate gradients are known to occur, and characterized the profiles of substrate, pH, oxygen, carbon dioxide, and temperature. The physically grounded steady-state axial diffusion equations with first- and zeroth-order kinetics yielded analytical solutions to the relevant variables. The results were compared with large-scale Escherichia coli and Saccharomyces cerevisiae experiments and simulations from the literature, and good agreement was found in substrate profiles. The analytical profiles obtained for dissolved oxygen, temperature, pH, and CO 2 were also consistent with the available data. Distribution functions for the substrate were defined, and efficiency factors for biomass growth and oxygen uptake rate were derived. In conclusion, this study demonstrated that axial diffusion equations can be used to model the effects of mixing and reaction on the relevant variables of typical large-scale fed-batch fermentations.
Enabling real-time monitoring and control of the biomanufacturing processes through product quality insights continues to be an area of focus in the biopharmaceutical industry. The goal is to manufacture products with the desired quality attributes. To realize this rigorous attribute-focused Quality by Design (QbD) approach, it’s critical to support the development of processes that consistently deliver high-quality products and facilitate product commercialization. Time delays associated with off-line analytical testing can limit the speed of process development. Thus, developing and deploying analytical technology is necessary to accelerate process development. In this study, we have developed the Micro Sequential Injection (µSI) process analyzer and the Automatic Assay Preparation Platform (A2P2) system. These innovations address the unmet need for an automatic, online, real-time sample acquisition and preparation platform system for in-process monitoring, control, and release of biopharmaceuticals. These systems can also be deployed in laboratory areas as an off-line analytical system and on the manufacturing floor to enable rapid testing and release of products manufactured in a GMP environment.
Enzymes that catalyze post-translational modifications of peptides and proteins (PTM-enzymes) – proteases, protein ligases, oxidoreductases, kinases, and other transferases - are foundational to our understanding of health and disease and empower applications in chemical biology, synthetic biology, and biomedicine. To fully harness the potential of PTM-enzymes, there is a critical need to decipher their enzymatic and biological mechanisms, develop molecules that can probe and reprogram them, and endow them with improved and novel functions. These objectives are contingent upon implementation of high-throughput functional screens and selections that interrogate large sequence libraries to isolate desired PTM-enzyme properties. This review discusses the principles of S. cerevisiae organelle sequestration to study and engineer PTM-enzymes. These include methods that modify yeast surface display and employ enzyme-mediated transcription activation to evolve the activity and substrate specificity of proteases and protein ligases. We also present a detailed discussion of yeast endoplasmic reticulum (ER) sequestration for the first time. Where appropriate, we highlight the major features and limitations of different systems, specifically how they can measure and control enzyme catalytic efficiencies. Taken together, yeast-based high-throughput sequestration approaches significantly lower the barrier to understanding how PTM-enzymes function and how to reprogram them.
Genome-scale metabolic models provide a valuable resource to study metabolism and cell physiology. These models are employed with approaches from the constraint-based modelling framework to predict metabolic and physiological phenotypes. The prediction performance of genome-scale metabolic models can be improved by including protein constraints. The resulting protein-constrained models consider data on turnover numbers (kcat) and facilitate the integration of protein abundances. In this systematic review, we present and discuss the current state-of-the-art regarding the estimation of kinetic parameters used in protein-constrained models. We also highlight how data-driven and constraint-based approaches can aid the estimation of turnover numbers and their usage in improving predictions of cellular phenotypes. Lastly, we identify standing challenges in protein-constraint metabolic models and provide a perspective regarding future approaches to improve the predictive performance.
Biomaterials are important tools for the reconstruction of organs and tissues, and silica is widely used in these reconstruction technologies. Hence, a lexical and content analysis was carried out on articles on application of silica in biomaterials, based on a search in the Capes’ Journals Portal using the keywords “silica, biomedical, application, scaffold”, between the years 2009 and 2019. The 110 selected articles were analyzed using the IRaMuTeQ software, Word Cloud and Descending Hierarchical Classification (DHC). The words: “silica, cell, nanosilica, bone, material, scaffold and application” appear prominently in the Word Cloud and DHC indicating four classes: (1) physical characterization, (2) biomedicine applications, (3) engineering applications and (4) compatibility characterization. Thus, the analysis of DHC and Word Cloud showed that the main ways of using silica are: mesoporous silica nanoparticles, amorphous silica, silica-based materials, nanofibers and silica hybrids, and the main biomaterials developed are scaffolds, grafts, aerogels, hydrogels, membranes and drug delivery systems.
Targeted gene knockdown has become one of the most powerful tools in molecular biology and holds substantial promise in therapeutic applications. While existing technologies such as siRNAs, CRISPRi, and ASOs effectively and specifically reduce gene expression, few can be used to first discover the genes that influence a particular phenotype and then directly transition to being used as oligonucleotide therapeutics. Thus, a tool that could help bridge the gap between target discovery and the development of therapeutic leads would benefit the scientific community. Here, we present hnRNPA1 recruiting oligonucleotides, or AROs, as single-stranded RNA (ssRNA) molecules that knockdown transcript levels of target genes. AROs target specific pre-mRNA transcripts via sequence homology and leverage the ubiquitous and abundant endogenous RNA-binding protein hnRNPA1 to degrade target transcripts. Using RT-qPCR, we show that AROs effectively knock down target genes when delivered via a plasmid and expressed using a Pol II promoter or when delivered directly as single-strand RNAs. Additionally, as proof of principle, we use a ssRNA ARO to knockdown KRT14 in squamous cell carcinoma and show reduced invasive potential. We believe AROs fill an important niche in the scientific toolbox by taking advantage of endogenous RNA binding machinery for RNA knockdowns.
Abstract Generally, investigations on nanomedicine involve conventional imaging techniques for obtaining static images on nanoparticle internalization at a single time point where various phases can be overlooked. In contrast, 3D live-cell imaging can be used for obtaining cellular retention of drugs at various phases, and cells can be followed for days. This article demonstrates the application of time-lapse microscopy in the investigation of Poly-L-lysine coated ZnO nanoparticle dynamics. In this work, a laser scanning confocal microscope has been employed to quantify the dynamics of internalization particles and reactive oxygen species generation (ROS) using volumetric imaging. Firstly, we show that simultaneous spatial mapping of nanoparticle uptake in MCF-7 cells and ROS in a single cell can be used to identify the interdependence between the accumulation of particles and ROS generation. Secondly, monitoring of ROS formation and cytotoxicity using the same imaging platform offers an advantage over monitoring these parameters using various instruments. Finally, the ability of the fluorescent particles in inducing a significant reduction in cell viability suggests its potential to be used as a therapeutic agent. The proposed framework opens up a new avenue of research for investigating mechanistic aspects of ZnO particle adsorption in vitro through long term imaging. Keywords: Fluorescent ZnO particle, Time-lapse microscopy, 3D Live-cell imaging, laser scanning confocal microscope, Reactive oxygen species
The global pandemic outbreak, SARS-COV-2, which causes COVID-19, has coerced numerous pharmaceutical companies to sprint for the vaccine and therapeutic biologics development. Most of the therapeutic biologics are common human IgG antibodies, which were identified by next-generation sequencing with the B cells from the convalescent patients in less than one-month post-infection. While the global public health emergency calls for medications urgently, it saves lives to expedite the clinical trials of biologics as much as possible, hence the biologics development strategies are unprecedentedly challenged. Since the advent of therapeutic biologics, transfection, and selection strategy has been continuously improving for developing more robust cell lines with greater productivity and efficiency. Next-generation sequencing (NGS) has also been implemented into cell bank testing for acceleration. These recent advances enable us to rethink and reshape the chemistry, manufacturing and controls (CMC) strategy against the pandemic outbreaks, to start supplying cGMP materials for the life-saving clinical trials as soon as possible. We elucidated an accelerated CMC workflow for biologics against pandemics, including using cGMP-compliant pool materials for Phase I clinical trials, selecting the final clone with similar product quality as Phase I materials for late-stage development and commercial production and matching product quality among different manufacturing stages.