Chemical modifications in therapeutic protein aggregates generated under different stress conditions

In this study, we characterized the chemical modifications in the monoclonal antibody (IgG(2)) aggregates generated under various conditions, including mechanical, chemical, and thermal stress treatment, to provide insight into the mechanism of protein aggregation and the types of aggregate produced by the different stresses. In a separate study, additional biophysical characterization was performed to arrange these aggregates into a classification system (Joubert, M. K., Luo, Q., Nashed-Samuel, Y., Wypych, J., and Narhi, L. O. (2011) J. Biol. Chem. 286, 25118-25133). Here, we report that different aggregates possessed different types and levels of chemical modification. For chemically treated samples, metal-catalyzed oxidation using copper showed site-specific oxidation of Met(246), His(304), and His(427) in the Fc portion of the antibody, which might be attributed to a putative copper-binding site. For the hydrogen peroxide-treated sample, in contrast, four solvent-exposed Met residues in the Fc portion were completely oxidized. Met and/or Trp oxidation was observed in the mechanically stressed samples, which is in agreement with the proposed model of protein interaction at the air-liquid interface. Heat treatment resulted in significant deamidation but almost no oxidation, which is consistent with thermally induced aggregates being generated by a different pathway, primarily by perturbing conformational stability. These results demonstrate that chemical modifications are present in protein aggregates; furthermore, the type, locations, and severity of the modifications depend on the specific conditions that generated the aggregates.     

Protein micelles from lipoxygenase 3

Heat-induced conformational changes in lipoxygenase 3 were characterized by differential scanning calorimetry. The positions of the observed transitions were sensitive to the composition of the buffer. In particular, lipoxygenase 3 heated in carbonate buffer at pH 8.0 formed large soluble aggregates. Variable-temperature circular dichroism revealed that the formation of the aggregates was not accompanied by the unfolding of the C-terminal domain, which is composed primarily of alpha-helix. The aggregates were investigated using size exclusion chromatography, native polyacrylamide gel electrophoresis, dynamic light scattering, and electron microscopy. The data were consistent with the formation of roughly spherical particles with an average hydrodynamic radius of 26 nm and an approximate composite molecular weight of 10,000,000 Da. To account for the formation of soluble aggregates from lipoxygenase 3, we propose that hydrophobic amino acid residues are exposed by unfolding of the N-terminal beta-barrel domain of the protein resulting in the formation of protein micelles with a hydrophilic surface composed of the C-terminal domains.     

Biological chemistry of immunomodulation by zwitterionic polysaccharides

Capsular polysaccharides isolated from pathogenic bacteria are comprised typically of many repeating units from one to eight or more monosaccharides in length. These polysaccharides stimulate the murine humoral immune system to elicit primarily IgM antibody responses. Studies conducted primarily in the mouse have characterized these polymers as T cell-independent antigens. These mouse studies and the relatively poor immunogenicity of polysaccharides in human hosts have led to the design of vaccines by coupling these polysaccharides to protein carriers to stimulate a T cell-dependent response. However, a newly described class of bacterial polysaccharides has been characterized that have the ability to modulate the cellular immune system. They are structurally diverse, but all share a zwitterionic charge motif that allows them to directly interact with T cells and antigen-presenting cells to initiate an immunomodulatory T cell response. These polymers, termed zwitterionic polysaccharides (ZPSs), elicit T cell-derived chemokines and cytokines that influence the immune response governing at least one classic host response to bacterial infection: abscess formation. This review will describe the biological and structural aspects of ZPSs that convey these activities.     

THE PROPERTIES OF SPECIFIC STAINS FOR ELECTRON MICROSCOPY PREPARED BY THE CONJUGATION OF ANTIBODY MOLECULES WITH FERRITIN

In order to take full advantage of recent developments in the electron microscopic examination of cellular ultrastructure and composition, it is necessary to develop specific electron stains capable of identifying and localizing a wide variety of macromolecular components of cells. To this end, antibody conjugates have been prepared by chemically coupling the highly electron-scattering ferritin molecule to antibody. Antigen-antibody precipitations with these ferritin-antibody conjugates have demonstrated that under the appropriate conditions they retain the specific binding properties of the antibody from which they are prepared. An electron microscopic study has been made of aggregates of tobacco mosaic virus and its ferritin-conjugated antibody. The aggregates were prepared in solution and then sprayed onto specimen screens. The electron micrographs reveal that the conjugate specifically attached to, and delineated, the virus rods. The chemistry, structure, and resolving power of the ferritin-antibody conjugates, the specificity of their reactions with homologous antigen, and the nature of the problems to be faced in application of these conjugates to the study of the internal antigens of cells are discussed.     

Expression of antibodies using single‐open reading frame vector design and polyprotein processing from mammalian cells

We describe a novel polyprotein precursor‐based approach to express antibodies from mammalian cells. Rather than expressing heavy and light chain proteins from separate expression units, the antibody heavy and light chains are contained in one single‐open reading frame (sORF) separated by an intein gene fused in frame. Inside mammalian cells this ORF is transcribed into a single mRNA, and translated into one polypeptide. The antibody heavy and light chains are separated posttranslationally, assembled into the functional antibody molecule, and secreted into culture medium. It is demonstrated that Pol I intein from P. horikoshii mediates protein splicing and cleavage reactions in mammalian cells, in the context of antibody heavy and light chain amino acid sequences. To allow the separation of antibody heavy chain, light chain, and the intein, we investigated a number of intein mutations designed to inhibit intein‐mediated splicing but preserve cleavage reactions. We have also designed constructs in which the signal peptide downstream from intein has altered hydrophobicity. The use of some of these mutant constructs resulted in more efficient antibody secretion, highlighting areas that can be further explored in improving such an expression system. An antibody secreted using one of the sORF constructs was characterized. This antibody has correct N‐terminal sequences for both of its heavy and light chains, correct heavy and light chain MW as well as intact MW as measured by mass spectrometry. Its affinity to antigen, as measured by surface plasmon resonance (SPR), is indistinguishable from that of the same antibody produced using conventional method. © 2009 American Institute of Chemical Engineers Biotechnol. Prog., 2009     

Prospects for advancing the understanding of complex biochemical systems

The application of mathematical theories to understanding the behaviour of complex biochemical systems is reviewed. Key aspects of behaviour are identified as the flux through particular pathways in a steady state, the nature and stability of dynamical states, and the thermodynamic properties of systems. The first of these is dealt primarily in theories of metabolic control, and metabolic control analysis (MCA) is an important example. The valid application of this theory is limited to steady-state systems, and the cases where the essential features of control can be derived from calibration experiments which perturb the state of the system by a sufficiently small amount from its operating point. In practice, time-dependent systems exist, it is not always possible to know a priori whether applied perturbations are sufficiently small, and important features of control may lie farther from the operating point than the application of the theory permits. The nature and stability of dynamical and thermodynamical states is beyond the scope of MCA. To understand the significance of these limitations fully, and to address the dynamical and thermodynamical properties, more complete theories are required. Non-linear systems theory offers the possibility of studying important questions regarding control of steady and dynamical states. It can also link to thermodynamic properties of the system including the energetic efficiency of particular pathways. However, its application requires a more detailed characterisation of the system under study. This extra detail may be an essential feature of the study of non-equilibrium states in general, and non-ideal pathways in particular. Progress requires considerably more widespread integration of theoretical and experimental approaches than currently exists.     

Visualization of aggregation of the Rnq1 prion domain and cross-seeding interactions with Sup35NM

Factors triggering the de novo appearance of prions are still poorly understood. In yeast, the appearance of one prion, [PSI(+)], is enhanced by the presence of another prion, [PIN(+)]. The [PSI(+)] and [PIN(+)] prion-forming proteins are, respectively, the translational termination factor Sup35 and the yet poorly characterized Rnq1 protein that is rich in glutamines and asparagines. The prion domain of Rnq1 (RnqPD) polymerizes more readily in vitro than the full-length protein. As is typical for amyloidogenic proteins, the reaction begins with a lag phase, followed by exponential growth. Seeding with pre-formed aggregates significantly shortens the lag. A generic antibody against pre-amyloid oligomer inhibits the unseeded but not the self-seeded reaction. As revealed by electron microscopy, RnqPD polymerizes predominantly into spherical species that eventually agglomerate. We observed infrequent fiber-like structures in samples taken at 4 h of polymerization, but in overnight samples SDS treatment was required to reveal fibers among agglomerates. Polymerization reactions in which RnqPD and the prion domain of Sup35 (Sup35NM) cross-seed each other proceeded with a shortened lag that only depends weakly on the protein concentration. Cross-seeded Sup35NM fibers appear to sprout from globular RnqPD aggregates as seen by electron microscopy. RnqPD spherical aggregates appear to associate with and, later occlude, Sup35NM seed fibers. Our kinetic and morphological analyses suggest that, upon cross-seeding, the aggregate provides the surface on which oligomers of the heterologous protein nucleate their subsequent amyloid formation.     

Oxidized human beta2-glycoprotein I: its impact on innate immune cells

Beta2-glycoprotein I (β2-GPI), an abundant 50 kDa plasma glycoprotein, is the most common target for antiphospholipid antibodies (aPLs). These autoantibodies are associated with thrombotic events in patients with anti-phospholipid antibody syndrome (APS) and systemic lupus erythematosus (SLE) and are proatherogenic. β2-GPI can also stimulate a vigorous adaptive cellular immune response in these patients. Although much is known about β2-GPI as a cofactor in autoimmune diseases, crucial information is still lacking to clarify why this abundant self plasma protein is the target of autoimmune responses. Throughout the years, a remarkable number of theories have been proposed to explain how the immune system recognises self. On the basis of a large variety of epidemiological, clinical and experimental evidence, it has been suggested that an unfortunate interplay of genetic susceptibility and environmental factors may play an important role in generating an abnormal immune response. Among the environmental factors, oxidative stress is one of the major events causing protein structural modifications, thus inducing the appearance of neo/cryptic epitopes of β2-GPI able to activate the immune system. In particular, oxidized β2-GPI is able to induce phenotypic and functional maturation of dendritic cells which represent the link between innate and adaptive immunity. Chronic activation of autoimmune reactions against this self protein modified by oxidative events may contribute to local and systemic inflammation, thus sustaining endothelial dysfunction in patients with APS, SLE and cardiovascular diseases. The role of oxidative stress in β2-GPI-mediated immune response is described in the light of our research experience and of relevant literature emerging in the field.     

Oxidation and reduction of cytochrome c bound to the phosphoprotein phosvitin

The oxidation-reduction reactions and structural characteristics of phosvitin-bound cytochrome c were examined at various ratios of cytochrome c to phosvitin. At binding ratios below half the maximum, the rate constants for the oxidation reactions with cytochrome c oxidase and ferricyanide and the rate constants for the reduction reactions with cytochrome b2 and ascorbate were low, but at higher ratios these rate constants gradually increased to that of free cytochrome c and, in particular, the rate constant for oxidation by cytochrome c oxidase was raised to two to three times that of the free form. This binding-ratio dependence of the rate constants for the oxidation and reduction reactions was different from that of the net charge of the cytochrome c-phosvitin complex, implying that the negative charges of phosvitin are unlikely to modulate the rates. In contrast, the broadening of the NMR signals for the heme and methionine-80 methyl groups and the conformational transition in the vicinity of the heme moiety on change from the native to the cyanide-bound or urea-denatured form of cytochrome c showed a similar binding-ratio dependence to the rate constants for the oxidation and reduction reactions. Since the conformation and electronic structure in the heme environment of ferric and ferrous cytochromes c were not changed significantly by binding to phosvitin, and since the binding strength of cytochrome c to phosvitin at binding ratios below half the maximum is different from that at higher ratios, these findings suggest that a difference in the movement of cytochrome c in its complex with phosvitin may modulate its oxidation-reduction reactions.     

Model immune complexes formed between monoclonal antibodies and a novel bivalent affinity label

Murine monoclonal anti-dinitrophenyl antibodies were produced by lymphocyte hybridomas and incubated with a new bivalent affinity label, bis-(dinitrofluorobenzene)-pimelic acid amide. Stable dimers were isolated from the resulting mixture of immune complexes in high purity. Experiments were performed to show that the immune complexes are covalently cross-linked through antibody active sites. Model dimers were also formed with varying ratios of monoclonal anti-hapten antibodies based on immunoglobulin isotype. The potential uses of these complexes for analyses of the biological activites of immune aggregates are discussed.     

Affinity precipitation of an antibody by ligand-modified phospholipids

A new method for the selective precipitation of proteins is applied to the isolation and purification of an antibody. Ligand-modified phospholipids (LMPs) are solubilized by the nonionic ethoxylated alcohol detergent, resulting in small (50 to 100 A) micellar aggregates of LMPs and surfactant. When introduced into protein solutions containing an antibody for which the LMP has specific affinity, the ligand binds to the protein. Hydrophobic interactions between phospholipid tail groups bound to the protein molecules result in an insoluble precipitate. Polyclonal and monoclonal antibiotin antibody (pABA and mABA) are shown to be selectively precipitated using ratios of dimyristoylphosphatidylethanolamidobiotin (DMPE-B) to ABA ranging from 1:1 to 19:1. The kinetics and yield of the precipitation achieve a maximum at a ratio of DMPE-B to ABA of approximately 7:1. The kinetics and magnitude of the turbidity change are modeled using the Mie theory of light scattering coupled with the smoluchowski theory of aggregation. The kinetics are shown to be enhanced significantly by the addition of salt. In particular, the addition of 0.5 M ammonium sulfate salt increases the rate of precipitation by more than an order of magnitude. It is demonstrated that pABA can be recovered with total activity yields of 60% to 70% from mixtures containing nonspecific lgG antibodies in very high purity (>99%). (c) 1994 John Wiley & Sons, Inc.     

The CanOE strategy: integrating genomic and metabolic contexts across multiple prokaryote genomes to find candidate genes for orphan enzymes

Of all biochemically characterized metabolic reactions formalized by the IUBMB, over one out of four have yet to be associated with a nucleic or protein sequence, i.e. are sequence-orphan enzymatic activities. Few bioinformatics annotation tools are able to propose candidate genes for such activities by exploiting context-dependent rather than sequence-dependent data, and none are readily accessible and propose result integration across multiple genomes. Here, we present CanOE (Candidate genes for Orphan Enzymes), a four-step bioinformatics strategy that proposes ranked candidate genes for sequence-orphan enzymatic activities (or orphan enzymes for short). The first step locates "genomic metabolons", i.e. groups of co-localized genes coding proteins catalyzing reactions linked by shared metabolites, in one genome at a time. These metabolons can be particularly helpful for aiding bioanalysts to visualize relevant metabolic data. In the second step, they are used to generate candidate associations between un-annotated genes and gene-less reactions. The third step integrates these gene-reaction associations over several genomes using gene families, and summarizes the strength of family-reaction associations by several scores. In the final step, these scores are used to rank members of gene families which are proposed for metabolic reactions. These associations are of particular interest when the metabolic reaction is a sequence-orphan enzymatic activity. Our strategy found over 60,000 genomic metabolons in more than 1,000 prokaryote organisms from the MicroScope platform, generating candidate genes for many metabolic reactions, of which more than 70 distinct orphan reactions. A computational validation of the approach is discussed. Finally, we present a case study on the anaerobic allantoin degradation pathway in Escherichia coli K-12.     

Multiple two-polymerase mechanisms in mammalian translesion DNA synthesis

The encounter of replication forks with DNA lesions may lead to fork arrest and/or the formation of single-stranded gaps. A major strategy to cope with these replication irregularities is translesion DNA replication (TLS), in which specialized error-prone DNA polymerases bypass the blocking lesions. Recent studies suggest that TLS across a particular DNA lesion may involve as many as four different TLS polymerases, acting in two-polymerase reactions in which insertion by a particular polymerase is followed by extension by another polymerase. Insertion determines the accuracy and mutagenic specificity of the TLS reaction, and is carried out by one of several polymerases such as polη, polκ or polι. In contrast, extension is carried out primarily by polζ. In cells from XPV patients, which are deficient in TLS across cyclobutane pyrimidine dimers (CPD) due to a deficiency in polη, TLS is carried out by at least two backup reactions each involving two polymerases: One reaction involves polκ and polζ, and the other polι and polζ. These mechanisms may also assist polη in normal cells under an excessive amount of UV lesions.     

Disulfide-bond formation in the transthyretin mutant Y114C prevents amyloid fibril formation in vivo and in vitro

The Y114C mutation in human transthyretin (TTR) is associated with a particular form of familial amyloidotic polyneuropathy. We show that vitreous aggregates ex vivo consist of either regular amyloid fibrils or disordered disulfide-linked precipitates that maintain the ability to bind Congo red. Furthermore, we demonstrate in vitro that the ATTR Y114C mutant exists in three forms: one unstable but nativelike tetrameric form, one highly aggregated form in which a network of disulfide bonds is formed, and one fibrillar form. The disulfide-linked aggregates and the fibrillar form of the mutant can be induced by heat induction under nonreduced and reduced conditions, respectively. Both forms are recognized by the amyloid specific antibody MAB(39-44). In a previous study, we have linked exposure of this epitope in TTR to a three-residue shift in beta-strand D. The X-ray crystallographic structure of reduced tetrameric ATTR Y114C shows a structure similar to that of the wild type but with a more buried position of Cys10 and with beta-mercaptoethanol associated with Cys114, verifying the strong tendency for this residue to form disulfide bonds. Combined with the ex vivo data, our in vitro findings suggest that ATTR Y114C can lead to disease either by forming regular unbranched amyloid fibrils or by forming disulfide-linked aggregates that maintain amyloid-like properties but are unable to form regular amyloid fibrils.     

Alzheimer's disease: synapses gone cold

Alzheimer's disease (AD) is a progressive neurodegenerative disease characterized by insidious cognitive decline and memory dysfunction. Synapse loss is the best pathological correlate of cognitive decline in AD and mounting evidence suggests that AD is primarily a disease of synaptic dysfunction. Soluble oligomeric forms of amyloid beta (Aβ), the peptide that aggregates to form senile plaques in the brain of AD patients, have been shown to be toxic to neuronal synapses both in vitro and in vivo. Aβ oligomers inhibit long-term potentiation (LTP) and facilitate long-term depression (LTD), electrophysiological correlates of memory formation. Furthermore, oligomeric Aβ has also been shown to induce synapse loss and cognitive impairment in animals. The molecular underpinnings of these observations are now being elucidated, and may provide clear therapeutic targets for effectively treating the disease. Here, we review recent findings concerning AD pathogenesis with a particular focus on how Aβ impacts synapses.     

Soluble aggregates of the amyloid-beta protein selectively stimulate permeability in human brain microvascular endothelial monolayers

Cerebral amyloid angiopathy associated with Alzheimer's disease is characterized by cerebrovascular deposition of the amyloid-beta protein (Abeta). Abeta elicits a number of morphological and biochemical alterations in the cerebral microvasculature, which culminate in hemorrhagic stroke. Among these changes, compromise of the blood-brain barrier has been described in Alzheimer's disease brain, transgenic animal models of Alzheimer's disease, and cell culture experiments. In the current study, presented data illustrates that isolated soluble Abeta(1-40) aggregates, but not unaggregated monomer or mature fibril, enhance permeability in human brain microvascular endothelial monolayers. Abeta(1-40)-induced changes in permeability are paralleled by both a decrease in transendothelial electrical resistance and a re-localization of the tight junction-associated protein zonula occludin-1 away from cell borders and into the cytoplasm. Small soluble Abeta(1-40) aggregates are confirmed to be the most potent stimulators of endothelial monolayer permeability by establishing an inverse relationship between average aggregate size and stimulated changes in diffusional permeability coefficients. These results support previous findings demonstrating that small soluble Abeta(1-40) aggregates are also primarily responsible for endothelial activation, suggesting that these same species may elicit other changes in the cerebrovasculature associated with cerebral amyloid angiopathy and Alzheimer's disease.     

Emerging roles for the ubiquitin-proteasome system and autophagy in pancreatic beta-cells

Protein degradation in eukaryotic cells is mediated primarily by the ubiquitin-proteasome system and autophagy. Turnover of protein aggregates and other cytoplasmic components, including organelles, is another function attributed to autophagy. The ubiquitin-proteasome system and autophagy are essential for normal cell function but under certain pathological conditions can be overwhelmed, which can lead to adverse effects in cells. In this review we will focus primarily on the insulin-producing pancreatic beta-cell. Pancreatic beta-cells respond to glucose levels by both producing and secreting insulin. The inability of beta-cells to secrete sufficient insulin is a major contributory factor in the development of type 2 diabetes. The aim of this review is to examine some of the crucial roles of the ubiquitin-proteasome system and autophagy in normal pancreatic beta-cell function and how these pathways may become dysfunctional under pathological conditions associated with metabolic syndromes.     

Interactions between glucocorticoid receptor-bearing chromatin and antireceptor antibody preparations

Recent evidence indicates that the control of gene expression by steroid hormones is mediated by hormone-receptor complexes bound at specific chromosomal locations. The isolation of these in vivo sites of binding would be useful in an analysis of the mechanism of this control. We have therefore examined the interaction between two different antiglucocorticoid receptor antibody preparations and a defined chromatin fraction containing bound glucocorticoid receptors in order to test the feasibility of this approach. Both antibody preparations, when attached to sepharose, removed nucleosome-bound and nucleosome free receptor from solution, indicating that chromatin-bound receptor was exposed and available for reaction with antibody. The bulk of the chromatin, not containing receptor, was mainly unaffected during these reactions, showing that the antibodies exhibited significant specificity. To determine whether the nucleosome-bound receptor remained attached to the nucleosome during reaction with antibody, studies using soluble antibody were performed. One of the antibodies caused a shift in the sedimentation rate of the nucleosome-bound receptor from 11S to 11.5S, suggesting that an intact ternary complex of antibody-receptor-nucleosome had formed. The other antibody produced various-sized aggregates of the free and bound receptors. Surprisingly, we found that one of the antibodies reacted strongly with free and nucleosome-bound estrogen receptors as well as glucocorticoid receptors. These studies suggest that an antibody preparation with appropriate characteristics should permit isolation of chromosomal receptor binding sites.     

Immunogenicity of self-associated aggregates and chemically cross-linked conjugates of the 42 kDa Plasmodium falciparum merozoite surface protein-1

Self-associated protein aggregates or cross-linked protein conjugates are, in general, more immunogenic than oligomeric or monomeric forms. In particular, the immunogenicity in mice of a recombinant malaria transmission blocking vaccine candidate, the ookinete specific Plasmodium falciparum 25 kDa protein (Pfs25), was increased more than 1000-fold when evaluated as a chemical cross-linked protein-protein conjugate as compared to a formulated monomer. Whether alternative approaches using protein complexes improve the immunogenicity of other recombinant malaria vaccine candidates is worth assessing. In this work, the immunogenicity of the recombinant 42 kDa processed form of the P. falciparum merozoite surface protein 1 (MSP1(42)) was evaluated as a self-associated, non-covalent aggregate and as a chemical cross-linked protein-protein conjugate to ExoProtein A, which is a recombinant detoxified form of Pseudomonas aeruginosa exotoxin A. MSP1(42) conjugates were prepared and characterized biochemically and biophysically to determine their molar mass in solution and stoichiometry, when relevant. The immunogenicity of the MSP1(42) self-associated aggregates, cross-linked chemical conjugates and monomers were compared in BALB/c mice after adsorption to aluminum hydroxide adjuvant, and in one instance in association with the TLR9 agonist CPG7909 with an aluminum hydroxide formulation. Antibody titers were assessed by ELISA. Unlike observations made for Pfs25, no significant enhancement in MSP1(42) specific antibody titers was observed for any conjugate as compared to the formulated monomer or dimer, except for the addition of the TLR9 agonist CPG7909. Clearly, enhancing the immunogenicity of a recombinant protein vaccine candidate by the formation of protein complexes must be established on an empirical basis.     

Influence of osmolarity and pH increase to achieve a reduction of monoclonal antibodies aggregates in a production process

Anti PSA monoclonal antibodies for diagnostic use were produced in an in vitro system. After purification using Protein G affinity chromatography a percentage of about 10% of antibody aggregates remained. The use of monoclonal antibodies containing aggregates as a capture antibody in a diagnostic kit reduces the performance of the test making it often unacceptable. The aggregates could be eliminated using gel filtration chromatography but, in that way, the final recovery of the whole production process was only about 50%. Aggregation is favoured when the working pH is near to the isoelectric point of the antibody. We varied the culture medium composition, modifying pH and osmolarity. We tested different values of pH and osmolarity: 7.1, 7.5, 8.0, 8.5 for pH, and 300, 340, 367, 395 mOsm/kg H2O for osmolarity. By modification of the cell culture medium we obtained a significant decrease of monoclonal antibody aggregates in the production cycle. In this way we achieved higher recovery rate and could avoid gel filtration polishing step. The experiments were performed in two stages: first in culture flasks changing one parameter in each experiment, and then in spinner bottle using the best conditions obtained in the first stage. During scale up we used the modifications achieved from the experiment showed in this paper in our production by hollow fibre bioreactor with positive results.