João J. Clemente

Improving expansion of pluripotent human embryonic stem cells in perfused bioreactors through oxygen control

The successful transfer of human embryonic stem cell (hESC) technology and cellular products into clinical and industrial applications needs to address issues of automation, standardization and the generation of relevant cell numbers of high quality. In this study, we combined microcarrier technology and controlled stirred tank bioreactors, to develop an efficient and scalable system for expansion of pluripotent hESCs. We demonstrate the importance of controlling pO(2) at 30% air saturation to improve hESCs growth. This concentration allowed for a higher energetic cell metabolism, increased growth rate and maximum cell concentration in contrast to 5% pO(2) where a shift to anaerobic metabolism was observed, decreasing cell expansion 3-fold. Importantly, the incorporation of an automated perfusion system in the bioreactor enhanced culture performance and allowed the continuous addition of small molecules assuring higher cell concentrations for a longer time period. The expanded hESCs retained their undifferentiated phenotype and pluripotency. Our results show, for the first time, that the use of controlled bioreactors is critical to ensure the production of high quality hESCs. When compared to the standard colony culture, our strategy improves the final yield of hESCs by 12-fold, providing a potential bioprocess to be transferred to clinical and industrial applications.

Optimization of the production of virus‐like particles in insect cells

In this work the maximal operational hydrodynamic conditions (agitation and aeration rate) that cause no adverse effect in Sf-9 cells growth in SF900II serum-free medium were determined. Shear stresses higher than 1 N m-2 and aeration rates higher than 0.04 vvm affect cell growth and when these conditions increase to 1.5 N m-2 and 0.11 vvm, cell growth is completely inhibited with significant cell morphology changes and a strong decrease in viability. Although the pO2 did not show a significant effect upon cell growth in the range from 10 to 50%, cell infection and specific productivity were dramatically affected. The production was optimal at a pO2 of 25% with decreases higher than 50% being observed when the pO2 decreased to 10 or increased to 50%. The maximum product quality, i.e., the percentage of product in the form of high molecular weight particles, is not coincident with maximum product titer. Although the highest Pr55gag particle titer was obtained at 96 hours post infection (hpi) and at pO2 of 25%, the best product quality (defined by gel filtration chromatography and Western immunoblot) was obtained at 48 hpi, independently of the pO2 used. The effect of overcritical conditions upon productivity was also studied. As obtained for cell growth, cell infection is affected by shear stresses above 1 N m-2 and by aeration rates higher than 0.04 vvm, with decreases in Pr55gag particle titer higher than 70%, even when the overcritical values are still far from the limit at which cell death occurs. The results obtained and the optimization strategy used allowed the maximization of the oxygen supply without damaging the cells, with important consequences on the scale-up of a production process involving this insect cell/baculovirus expression system.

A procedure for high-yield spore production by Bacillus subtilis

Bacillus subtilis spores have a number of potential applications, which include their use as probiotics and competitive exclusion agents to control zoonotic pathogens in animal production. The effect of cultivation conditions on Bacillus subtilis growth and sporulation was investigated in batch bioreactions performed at a 2‐L scale. Studies of the cultivation conditions (pH, dissolved oxygen concentration, and media composition) led to an increase of the maximum concentration of vegetative cell from 2.6 × 109 to 2.2 × 1010 cells mL‐1 and the spore concentration from 4.2 × 108 to 5.6 × 109 spores mL‐1. A fed‐batch bioprocess was developed with the addition of a nutrient feeding solution using an exponential feeding profile obtained from the mass balance equations. Using the developed feeding profile, starting at the middle of the exponential growth phase and finishing in the late exponential phase, an increase of the maximum vegetative cell concentration and spore concentration up to 3.6 × 1010 cells mL‐1 and 7.4 × 109 spores mL‐1, respectively, was obtained. Using the developed fed‐batch bioreaction a 14‐fold increase in the concentration of the vegetative cells was achieved. Moreover, the efficiency of sporulation under fed‐batch bioreaction was 21%, which permitted a 19‐fold increase in the final spore concentration, to a final value of 7.4 × 109 spores mL‐1. This represents a 3‐fold increase relative to the highest reported value for Bacillus subtilis spore production.

Process development of a recombinant antibody/interleukin-2 fusion protein expressed in protein-free medium by BHK cells.

The production, purification and stability of quality (in terms of integrity and glycosylation) of an antibody/interleukin-2 fusion protein with potential application in tumour-targeted therapy expressed in BHK21 cells are described. Consistency of the product throughout time was determined by analysis of glycosylation of the fusion protein using MALDI-TOF mass spectroscopy and HPAEC-PAD combined with product integrity studies by SDS-PAGE and Western blotting. These investigations showed consistent expression in terms of integrity and of three major oligosaccharide structures of the fusion protein after 62 generations. The data obtained at this stage indicated the suitability of the cell line for production purposes. Different approaches for the production of this protein were subsequently carried out. The relative productivity of the recombinant fusion protein and general performance of the cells in two different protein-free medium (PFM) culture systems, continuous chemostat and continuous perfusion using a Centritech centrifuge as a cell retention device, were studied. The results indicate that the chemostat culture resulted in more stable and controllable nutrient environment, which could indicate better product consistency, in accordance with what has been observed under serum-containing conditions, in relation to the perfusion culture. Finally, product obtained from the chemostat culture was analysed and purified. The purification process was optimised with an increase in the overall yield from 38 to 70% being obtained, a significant improvement with important consequences for the implementation of an industrial-scale culture system. In conclusion, it was possible to produce and purify the recombinant antibody/interleukin-2 fusion protein assuring the quality and stability of the product in terms of integrity and glycosylation. Therefore, a candidate production process was established.

Merging bioreactor technology with 3D hepatocyte-fibroblast culturing approaches: Improved in vitro models for toxicological applications

During the last years an increasing number of in vitro models have been developed for drug screening and toxicity testing. Primary cultures of hepatocytes are, by far, the model of choice for those high-throughput studies but their spontaneous dedifferentiation after some time in culture hinders long-term studies. Thus, novel cell culture systems allowing extended hepatocyte maintenance and more predictive long term in vitro studies are required. It has been shown that hepatocytes functionality can be improved and extended in time when cultured as 3D-cell aggregates in environmental controlled stirred bioreactors. In this work, aiming at further improving hepatocytes functionality in such 3D cellular structures, co-cultures with fibroblasts were performed. An inoculum concentration of 1.2×10(5) cell/mL and a 1:2 hepatocyte:mouse embryonic fibroblast ratio allowed to improve significantly the albumin secretion rate and both ECOD (phase I) and UGT (phase II) enzymatic activities in 3D co-cultures, as compared to the routinely used 2D hepatocyte monocultures. Significant improvements were also observed in relation to 3D monocultures of hepatocytes. Furthermore, hepatocytes were able to respond to the addition of beta-Naphtoflavone by increasing ECOD activity showing CYP1A inducibility. The dependence of CYP activity on oxygen concentration was also observed. In summary, the improved hepatocyte specific functions during long term incubation of 3D co-cultures of hepatocytes with fibroblasts indicate that this system is a promising in vitro model for long term toxicological studies.

Influence of environmental factors on lipase production by Lactobacillus plantarum

Abstract A strain of Lactobacillus plantarum, DSMZ 12028 (Deutsch Sammlung von Mikroorganismen und Zellkulturen), isolated from a Portuguese dry fermented sausage, “chouriço”, was found to produce true lipase, producing free fatty acids from triolein (olive oil). This enzymatic activity was found in whole cells, but was negligible in comparison to lipolytic activity in culture supernatant. Therefore, only extracellular activity was studied. The effect of pH, temperature and glucose concentration on extracellular lipase production was studied in continuously stirred tank reactors, the first time this technology has been used to study the production of this enzyme in lactobacilli. Maximum lipase production was achieved at a pH of 5.5 and 30 °C and was kept at a significant level over a wide range of dilution rates (0.05–0.4 h−1); the production of lipase was still significant for low pH values, temperature and glucose concentration, conditions that are close to the ones present during chouriço ripening. The effect of glucose concentration was also studied in a batch system. The control of lipase production was found to be related both to glucose concentration in the medium and to the growth rate/dilution rate. Glucose concentration was found to be important for fast lipase production, although it did not influence the maximum lipase activity reached in a batch culture.

Bioprocess iterative batch-to-batch optimization based on hybrid parametric/nonparametric models.

This paper presents a novel method for iterative batch‐to‐batch dynamic optimization of bioprocesses. The relationship between process performance and control inputs is established by means of hybrid grey‐box models combining parametric and nonparametric structures. The bioreactor dynamics are defined by material balance equations, whereas the cell population subsystem is represented by an adjustable mixture of nonparametric and parametric models. Thus optimizations are possible without detailed mechanistic knowledge concerning the biological system. A clustering technique is used to supervise the reliability of the nonparametric subsystem during the optimization. Whenever the nonparametric outputs are unreliable, the objective function is penalized. The technique was evaluated with three simulation case studies. The overall results suggest that the convergence to the optimal process performance may be achieved after a small number of batches. The model unreliability risk constraint along with sampling scheduling are crucial to minimize the experimental effort required to attain a given process performance. In general terms, it may be concluded that the proposed method broadens the application of the hybrid parametric/nonparametric modeling technique to “newer” processes with higher potential for optimization.

Predicting sporulation events in a bioreactor using an electronic nose

An electronic nose (EN) based on a non‐ specific multi‐sensor array was used to accurately estimate sporulation events and the spore concentration of Bacillus subtilis cultures. The array included 6 metal oxide sensors (MOS), 10 metal oxide semiconductor field effect transistors (MOSFET), one CO2 infrared sensor and one humidity sensor. The EN was used to monitor the gas emissions from B. subtilis bioreactions during both batch and fed‐batch operation. The signal pattern produced by the sensors was evaluated by principal component analysis (PCA) and training cultivations were used to build a model. The arc length of the PCA trajectories was successfully correlated to the off‐line spore count; a strong linear correlation (R2 = 0.992) between the numerical integration of the curves and the measured spore concentration was established. The fast responses of the sensors in combination with the robust correlation with the off‐line determination of spore concentration establish this EN device as a convenient tool for monitoring sporulation events in bioprocesses. Biotechnol. Bioeng. 2008;101: 545–552.

Analysis of culture media screening data by projection to latent pathways: The case of Pichia pastoris X-33

Cell culture media formulations contain hundreds of individual components in water solutions which have complex interactions with metabolic pathways. The currently used statistical design methods are empirical and very limited to explore such a large design space. In a previous work we developed a computational method called projection to latent pathways (PLP), which was conceived to maximize covariance between envirome and fluxome data under the constraint of metabolic network elementary flux modes (EFM). More specifically, PLP identifies a minimal set of EFMs (i.e., pathways) with the highest possible correlation with envirome and fluxome measurements. In this paper we extend the concept for the analysis of culture media screening data to investigate how culture medium components up-regulate or down-regulate key metabolic pathways. A Pichia pastoris X-33 strain was cultivated in 26 shake flask experiments with variations in trace elements concentrations and basal medium dilution, based on the standard BSM+PTM1 medium. PLP identified 3 EFMs (growth, maintenance and by-product formation) describing 98.8% of the variance in observed fluxes. Furthermore, PLP presented an overall predictive power comparable to that of PLS regression. Our results show iron and manganese at concentrations close to the PTM1 standard inhibit overall metabolic activity, while the main salts concentration (BSM) affected mainly energy expenditures for cellular maintenance.

Design of Pathway-Level Bioprocess Monitoring and Control Strategies Supported by Metabolic Networks

In this chapter we explore the basic tools for the design of bioprocess monitoring, optimization, and control algorithms that incorporate a priori knowledge of metabolic networks. The main advantage is that this ultimately enables the targeting of intracellular control variables such as metabolic reactions or metabolic pathways directly linked with productivity and product quality. We analyze in particular design methods that target elementary modes of metabolic networks. The topics covered include the analysis of the structure of metabolic networks, computation and reduction of elementary modes, measurement methods for the envirome, envirome-guided metabolic reconstruction, and macroscopic dynamic modeling and control. These topics are illustrated with applications to a cultivation process of a recombinant Pichia pastoris X33 strain expressing a single-chain antibody fragment (scFv).