Cell growth

W T. Godbey , in Biotechnology and its Applications (Second Edition), 2022

5.2.7 Be aware of the lag phase

Consider a case where a batch cell culture has been incubated for too long, and a small aliquot of cells is transferred from this plateau phase culture and placed in a new bioreactor with fresh culture medium. Referring to the graph in Figure 5.8, we see that the new growth curve has the same elements we've already covered: it has the same general shape and the same phases as before but a new y 0 and a different lag phase length. The length of the lag phase is related to how different the new culture environment is from the old one. For instance, taking cells from the end of lag phase or the very beginning of the log phase and starting a new culture with them should yield a relatively short lag phase for the new culture because the difference between the media should be very small. However, starting a new culture with cells from a plateau phase culture should yield a longer lag phase because the old and new culture media are very different. The cells may have to turn on or off many genes because they are being transferred from a nutrient-depleted environment to one that is rich in resources. The transferred cells first must detect the new surroundings, then begin transcribing genes that are appropriate for the new surroundings. Because the cells will begin growing and dividing again, they will have to produce enzymes that will aid with the production of dNTPs, DNA polymerase, transcription factors, cytoskeletal proteins, and so on. The further away the parent culture was from the lag phase on its own growth curve (every culture has its own growth curve), the longer the next lag phase will be. That's why, instead of taking our culture from late in the plateau phase (which might make sense because there will be a maximal number of cells), we might choose to pass the cells to a new culture during late log phase to shorten the time needed to bring them back into exponential growth. Again, this discussion is for cells grown in batch.

Figure 5.8. Growth curves to consider. Starting with a parent culture of Escherichia coli cells (black curve), if we were to take a standard aliquot at the point indicated by the black arrow and start a new culture, the new culture would have a growth curve like the black line. Taking the standard aliquot at an earlier time (blue arrow) would yield a new culture with a growth curve like the blue curve. Note the lower starting concentration, shorter lag phase, and slightly longer plateau phase. Taking the standard aliquot at a later time (red arrow) would yield a new culture with a growth curve like the red curve. Note the higher starting concentration, longer lag phase, and slightly shorter plateau phase. The maximum number of cells in each culture should be the same because the results of contact inhibition should be unchanged for these cells. (Note that the curves do not represent actual values but are drawn to illustrate the principles mentioned here.)

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780128177266000058

Plant Tissue Culture

Saurabh Bhatia , in Modern Applications of Plant Biotechnology in Pharmaceutical Sciences, 2015

2.2.2.2 Cell Suspension Cultures

Haberlandt made the first attempt to isolate single cells from leaves of flowering plants [158]. Muir in 1958 clearly showed that plant cells can be grown in liquid (suspension) cultures similar to microorganisms [159]. However, plant cells individually in a population of cultured cells invariably show cytogenetical and metabolic variations depending upon the stage of the growth cycle and culture conditions. This type of variability is called spatial heterogeneity [160]. Suspension culture is the culturing of isolated cells in liquid media. Cells can be isolated from in vivo plant material, either by mechanical means or through enzymatic digestion, and form callus induced from any explants. Cell suspension cultures can be initiated from virtually any part of the plant, just as callus culture. It is the most common way to initiate suspension cultures from callus already growing in culture. Broadly callus cultures are divided into two different categories: friable and compact. Cells are densely aggregated in compact callus whereas cells are loosely associated in friable callus. Based on the type of callus (friable callus or soft callus requires less force to break apart easily than compact callus) suspension cultures are produced. These types of callus offer the inoculums to form cell suspension cultures. For initiating cell suspensions relatively large inoculums should be used so that the released concentration increases quickly. Such types of cells represent a lower level of organization than tissue or callus culture and are often called cell cultures. This liquid medium is typically similar in composition to the one on which the callus is grown. However, adjustments may have to be made in the concentration of hormone and inorganic salts. Agitation is required for suspension cultures for three purposes: it serves to break up the cell aggregates; it maintains a uniform distribution of cells of various sizes and shapes; and cell clumps in the medium provide gaseous exchange for the cells to sustain cell respiration in the liquid medium. As cell division proceeds, the inoculums break up and shed new cell clusters, which fracture again to give individual cells and other small groups. Due to the natural tendency of plant cells to adhere together, it is not always possible to grow a suspension of single cells only. Different types of suspension cultures are highlighted in Fig. 2.7.

Figure 2.7. Types of suspension culture.

2.2.2.2.1 Batch cultures

Batch culture is a closed culture system that contains limited amounts of nutrients. In batch culture cells grow in a finite volume of liquid medium and are usually maintained in conical flasks on orbital shakers at a speed of 80–120 rpm. There are many types of batch culture: slowly rotating culture, shake culture, spinning culture, and stirred culture.

2.2.2.2.2 Continuous cultures

Cultures that require a continuous supply of the cell suspension or the product in the medium are known as continuous cultures. This system is maintained in a steady state for prolonged periods by draining out the used liquid medium and adding fresh medium to stabilize the physiological state of growing cells. The growth medium is designed in such a way that one of the nutrients is in limited quantity therefore during exponential growth as the nutrient is exhausted the growth will stop. However, before the nutrient is exhausted the fresh medium with limited nutrient is added. Due to the continuous flow of the nutrients and steady state of the cells' nutrients depletion does not occur. A wide range of bioreactor configurations and sizes have been designed for continuous cultures depending on the variety of plant cells [161,162]. There are two types of continuous cultures: (i) closed continuous culture – in this system cells are separated from the drained medium and added back to suspension culture and (ii) open continuous culture – in this system addition of the medium is accompanied by harvest of an equal volume of suspension culture.

2.2.2.2.3 Suspension culture techniques

Suspension cultures of higher plant cells in synthetic media were demonstrated by Torrey [163]. In 1965, Earle and Torrey suggested the defined media for the cells isolated from Convolvulus. Well colonies were formed on his suggested media. Later on several researchers defined different media for different plant cell cultures [164]. For initiating cell cultures at low inoculum density a conditioned medium is used. Culture medium for cell suspension and its conditioning was demonstrated by Torres (Fig. 2.8) [165]. This involves the separation of high-density cell culture from a low-density culture medium by a barrier that permits the diffusion of solutes and air. There are so many techniques used for the culturing of cell culture. Among them the oldest and most frequently utilized technique is Bergmann's cell plating technique (Fig. 2.8). In this technique free cells are suspended in liquid medium (if cell aggregates are present, the culture is filtered), and a culture medium with agar (0.6–1%) is cooled and maintained at 35°C in a water bath [166]. An equal volume of these liquid and agar media are mixed and rapidly spread in a Petri dish, so that cells are evenly distributed in a thin layer, after solidification. The Petri dishes are sealed with parafilm and examined with an inverted microscope to mark single cells (marking is done on the outer surface of the dish). The plates are incubated in the dark at 25°C and cell colonies developing from marked single cells are used to obtain single cell cultures. Various other methods (e.g., filter paper raft-nurse technique; microchamber technique, bioreactors, etc.) have also been developed to grow individual cells. The microchamber technique was first attempted by De Ropp, then successfully accomplished by Jones et al. In this technique a drop of medium containing a single cell is placed over a microscopic slide and covered with three different cover slips in such a fashion that a microchamber is formed [167,168]. Before placing each cover slip, mineral oil is poured on each occasion aseptically. The formed microchamber protects the plant cell from water loss but permits the gaseous exchange (Fig. 2.8). These cultured cells can be preserved by cryopreservation. Mustafa et al. demonstrated the two-step (controlled rate) freezing technique, also known as the slow (equilibrium) freezing method for long-term storage, which has been applied successfully to a wide range of plant cell suspension cultures [169].

Figure 2.8. (a) Conditioning apparatus for low-density cell culture medium [163]; (b) Bergmann's cell plating technique for culturing single cells [166]; (c) microchamber technique for cell plating [168].

2.2.2.2.4 Pattern of growth

Cells in suspension can exhibit much higher rates of cell division than those in callus culture. Suspension cultures when maintained under controlled conditions of light, temperature, and aeration follow a predictable pattern of growth curve (Fig. 2.9)

Figure 2.9. Growth curve showing different phases of growth.
•

Lag phase: The culture first passes through a lag phase in which there is little growth. The lag phase is the period when the cells adjust to the replenished supplies of nutrients and undertake all the necessary synthesis prior to cell division.

•

Log phase: The cultures then pass through the logarithm phase or exponential phase of growth in which cells divide very rapidly, causing a logarithmic increase in cell number. Under optimum conditions, cell numbers double every 20–50 h, depending upon their species.

•

The culture passes through a further period of rapid cell division that results in a linear increase in number, slowing at the phase as some nutrients become limiting.

•

Stationary phase: The cultures then reach a stationary phase, when the rate of cell division within the culture decreases, and the cell number is stabilized and growth finally halts. As nutrients are depleted, some of the cells of the culture begin to show senescent characteristics and a low level of cell division will maintain cell numbers. If the cells are left in the stationary phase too long, they will die.

2.2.2.2.5 Application of cell culture

For production of secondary metabolites (discussed in Chapter 7), mutant selection (discussed in Chapter 11), biotransformation (discussed in Chapter 7), somatic embryogenesis (discussed in Chapter 6), immobilization (discussed in Chapter 7), commercial production of cell mass and induction of mutation and genetic manipulation of plant cells, studying the effect of different chemicals, for obtaining individual isolated cells used for protoplast production, commercial production of cell mass, single cell protein hybrids or cybrids, to develop the single cell line or cell clone and easily preserved by using cryopreservation technique than other tissues or organ of the plant.

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780128022214000029

Bacterial and Yeast Cultures – Process Characteristics, Products, and Applications

Wei-Cho Huang , I-Ching Tang , in Bioprocessing for Value-Added Products from Renewable Resources, 2007

5.1 Culture characteristics affecting fermentation

The bacterial or yeast culture used in a fermentation process would determine the productivity, yield, and purity, which are also dependent on the operating conditions. The media composition (mainly carbon and nitrogen sources) and fermentation conditions (pH, temperature, mixing, aeration, etc.) are the common factors to be considered and need to be optimized based on the culture used in the fermentation. Furthermore, process design, including the reactor operation mode, down-stream processing, and even waste treatments, is also highly dependent on the culture. It is thus very important to select suitable cultures and processing conditions for economical industrial production. The selection of the culture for production of ethanol and lactic acid will be discussed as two examples.

As shown in Tables 6 and 7, each culture has its own pros and cons when used in the fermentation. The final selection thus will be based mainly on the economical factors, including raw material costs, productivity, yield, recovery costs, and waste disposal. For example, S. cerevisiae is widely used in industrial ethanol fermentation even though Zymomonas mobilis can produce ethanol from glucose at a higher productivity and yield [141, 142]. This is because the yeast cell is hardy and easy to separate from the fermentation broth. In order to use xylose for ethanol production, recombinant strains of S. cerevisiae, Z. mobilis, and E. coli have been developed. Although E. coli appears to be a better organism in simultaneously converting glucose and xylose to ethanol, its low ethanol tolerance, neutral pH for growth, and disposal after fermentation are unfavorable to its industrial application [28]. Yeast biomass generated in ethanol fermentation is used in animal feed and thus does not pose a disposal problem.

Table 6. Comparison of ethanol fermentations by bacterial and yeast cultures

G(-) Bacteria Yeast G(+) Bacteria
Species E. coli S. cerevisiae Zymomonas mobilis
Substrates Glucose, xylose Glucose, sucrose, xylose Glucose, fructose, sucrose; xylose
Medium Simple medium Simple medium Simple medium
Culture pH pH 6∼8 pH 5 pH 7
Product yield 0.46 g/g glucose 0.47 g/g glucose 0.49 g/g glucose
0.46 g/g xylose 0.43 g/g xylose nil from xylose
Product concentration ∼50 g/L 120 g/L from glucose 120 g/L from glucose
Productivity 0.83 g/L.h >1.4 g/L·h >2 g/L·h with glucose

Table 7. Comparison of lactic acid fermentations by bacterial, yeast, and filamentous fungal cultures

Bacteria Yeast Filamentous fungi
Species Lactobacillus spp. Kluyveromyces lactis Rhizopus oryzae
Substrates Glucose, lactose, sucrose; can't use starch Glucose, lactose; can't use starch Glucose, starch, xylose
Medium Require complex Relative simple Simple medium with
growth nutrients industrial media only trace minerals
Growth conditions Anaerobic, pH >5 Anaerobic, pH 4.5 Aerobic, pH > 4
Products Mixtures of L(+) and D(-)-lactic acids Pure L(+)-lactic acid Pure L(+)-lactic acid
Product yield 0.85 ∼ 0.95 g/g glucose ∼0.85 g/g glucose ∼0.85 g/g glucose
Product concentration Up to 150 g/L 60–109 g/L Up to 120 g/L
Productivity as high as 60 g/L·h 0.12–0.91 g/L·h Up to 6 g/L·h

Food and pharmaceutical grades of lactic acid are mainly produced by LAB, such as Lactobacillus spp. because of the high productivity and yield. However, industrial LAB strains usually produce a mixture of l-(+) and D-(−)-lactic acids, which is not suitable for the synthesis of polylactic acid. LAB also require a complex medium for their growth. Tate and Lyle developed a recombinant yeast Kluyveromyces lactis that can produce pure L-(+)-lactic acid from glucose in a simple medium but has a lower productivity and yield. The filamentous fungus Rhizopus oryzae also can produce pure L-(+)-lactic acid from glucose, starch, and xylose in a simple medium, but its growth requires aeration and its filamentous morphology can be difficult for scale up [143].

5.1.1 Media formulation and feeding mode

Culture media contain all the necessary nutrients to maintain cell growth and to generate products. Nutrients include carbon, nitrogen, oxygen, hydrogen, sulfur, phosphorus, trace elements, vitamins, growth factors, and metabolic precursors. Anti-foaming agent and buffering chemicals also may be added in the medium. Media formulation is determined by the nature of the desired fermentation products and the culture used. Two types of growth media are usually used in fermentation: synthetic (with a well defined composition) and complex media. Defined media have specific amounts of pure chemical compounds and an identifiable chemical composition. Complex (enriched) media contain natural compounds whose precise chemical composition is not known. For example, a medium containing yeast extract, peptone, molasses, or corn steep liquor is a complex medium, which provides necessary nutrients and generates higher cell yields than defined media. However, complex media can vary from one batch to another, and thus affect the process reproducibility; defined media are more reproducible, giving the operator better control of the fermentation, but are more expensive than complex media. For economic reasons, complex media are widely used in industrial fermentations. Glucose and sucrose are often used as the carbon source, but they may be supplied in an industrial substrate such as molasses, which provide not only the carbon source but also nitrogenous substances, vitamins, and trace elements. The required nitrogen source is usually supplied in yeast extract or corn steep liquor, an inexpensive industrial byproduct from corn refining. These materials constitute an excellent source of nitrogen as well as other growth factors. For some bacteria, urea and inorganic nitrogenous compounds, such as ammonia and ammonium sulfate, also have been used.

The choice of the carbon source, which is usually the major raw material cost, is dependent on the cultures. For example, amylase-producing microbes can utilize starch and dextrins, instead of the more expensive glucose. Genetically modified yeasts can use xylose, which is a major sugar component in hemicellulose. Kluyveromyces and Torulopsis spp. can consume the lactose in cheese whey as a cheap carbon source. Yarrowia lipolytica is able to degrade lipids, proteins and n-paraffins as sole carbon source. Candida can assimilate n-alkanes and fatty acids as carbon sources. Methylotrophic bacteria, such as Methylobacterium extorquens, grow on methanol utilizing the serine cycle for carbon assimilation. Some yeasts, including Candida, Hansenula, Pichia, and Torulopsis, also can use methanol as the carbon source.

Batch, fed-batch, and continuous cultures are three common ways to grow microorganisms. In batch cultures, cells are initially inoculated into a fresh medium and no further nutrient added until the target product is produced. In fed-batch cultures, growth medium is added at various intervals, while effluent is removed discontinuously. High cell density is usually attained in fed-batch cultures, since nutrients are added as required to maintain higher cell growth and to prolong the growth phase of the fermentation process. In continuous cultures, fresh growth medium is continually added throughout the whole process, and cells and spent medium are removed simultaneously. By and large, growth and uniform product formation are maintained for a longer period in continuous cultures. Although batch cultures are the most used bioprocesses, a fed-batch system combined with the features of continuous culture and batch growth is also widely used in commercial plants.

5.1.2 Cell density and immobilized cell cultures

The reactor productivity is usually proportional to the viable cell density in the fermenter. Cell recycling has been used to increase the cell density and product concentration of continuous cultures. It can also reduce the formation of inhibitory end-products resulting from lower concentrations of substrates due to the higher dilution rate applied. Filtration, centrifugation, sedimentation, and immobilization are the representative methods used to retain a high cell density. In addition, high cell density fermentations can be achieved using fed-batch technology, as it overcomes substrate and product inhibition. Cell immobilization is the most efficient among these techniques and contributes to high cell growth rates and long-term stability in a bioreactor. For example, butanol production in ABE fermentation can be improved by cell recycling and immobilization to increase cell density and reactor productivity [144–145].

Immobilized cell cultures have many advantages over suspension cultures for large-scale fermentation. They provide high cell density, eliminate expensive cell recovery, alleviate cell washout at high dilution rates, and protect against shear damage. Thus, cell immobilization enhances productivity in continuous processes. The adsorption of cells onto inert support surfaces has been widely used for cell immobilization (readers are referred to Chapter 14). For example, the fibrous bed bioreactor (FBB) improved several organic acid fermentations with significantly enhanced productivity, yield, and product concentration [146–147]. The fibrous bed allows for good multiphase flows and provides renewable surfaces for cell immobilization [148], resulting in a high cell density (see Fig. 3). FBBs provide efficient, continuous operation with high cell density and economic downstream processing in large-scale fermentations [149].

Fig. 3. SEM micrographs of C. acetobutylicum cells immobilized on fibers in a FBB. (Magnification: A: 500x; B: 2500x; C: 3500x; D: 5000x)

5.1.3 Metabolism and metabolic engineering

Many industrial applications exploit the specific capability of microorganisms to make a variety of products. Increasing the productivity and yield of certain primary or secondary metabolites has become the objective of many biotechnologists [150]. To explore the full industrial potential of bacterial and yeast cells, it is necessary to understand their growth and metabolic pathways, which are linked to the successful commercial exploitation of fermentation products.

Bacterial and yeast cells produce various products as a consequence of different metabolisms. However, all cells generate energy to drive vital functions via catabolism and synthesize biological compounds through anabolic pathways. Almost all cells share similar metabolic pathways, which perform enzymatic reactions to transform substrates into end products (amino acids, lipids, or polysaccharides). In general, aerobic catabolism consists of the Embden-Meyerhof-Parnas (EMP) pathway: fermentation of glucose to pyruvate, the tricarboxylic acid (TCA) cycle, which converts pyruvate to CO2 and NADH, and the electron transport chain, which forms ATP by electron transfer from NADH. The hexose-monophosphate (HMP) pathway, which can be shuttled into glycolysis, converts glucoses-6-phosphate into carbon skeletons and reducing power for direct use in biosynthesis. Entner-Doudoroff (ED) pathway converts glucose to pyruvate and glyceraldehyde-3 phosphate by producing 6-phosphogluconate and then dehydrating it.

Recently, metabolic engineering techniques have been applied to enhance the production of industrial bio-based products. Modified genes can lead microorganisms to new metabolic pathways for novel products or achieve higher efficiencies in metabolite overproduction through alterations in metabolic flux distribution. Recent progress in metabolic engineering has provided many significant applications for engineered strains and will continue to expand into new areas of applications as more and more genes become available. These metabolically engineered cells either overexpress controlling genes (or key enzyme) or block undesired metabolic pathways to overproduce value-added products. Therefore, an increasing number of microorganisms will become more amenable to genetic manipulation and facilitate the design and control of metabolic pathways.

Metabolic modeling aids quantitative study of the metabolism and provides new insights into microorganisms. One of the most popular methods is metabolic flux analysis (MFA), which provides a measure of the change in overall cellular functions and metabolic processes [151]. MFA measures the inputs and outputs of a cell and uses knowledge of the metabolic pathways to calculate the fluxes through these pathways. Stoichiometric mass balance is a popular and readily applied method to determine metabolic flux distribution in the central metabolism under pseudo-steady-state assumptions. It requires neither enzymatic kinetics nor expensive equipment and statistical calculations for isotope tracers but provides significant metabolic information [43]. A metabolic network with carbon fluxes provides a clear picture of carbon distribution or in vivo fluctuations that affect yield and productivity. For instance, flux analysis of yeast described ATP requirements for biomass synthesis and intermediate metabolite transportation in carbon-limited chemostat cultures and identified the critical pathway in the metabolism of recombinant S. cerevisiae for the production of ethanol [11]. These metabolic engineering tools greatly minimize experimental efforts for process optimization and develop effective strains for reducing production costs. More details about metabolic engineering can be found in Chapter 4.

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780444521149500098

Volume 3

B.K. Lavine , in Comprehensive Chemometrics, 2009

3.19.4.1 Cystic Fibrosis Heterozygotes versus Presumed Normal Donors

The first study 18–21 involves the application of pyrolysis gas liquid chromatography and supervised pattern recognition methods to the problem of identifying carriers of the cystic fibrosis (CF) defect. The biological samples used in this experiment were cultured skin fibroblasts grown from 24 samples obtained from parents of children with CF (4 males and 20 females) and from 24 presumed normal donors (16 males and 8 females). The cells were cultured in modified Eagle's minimum essential medium supplemented with 15% fetal bovine serum and gentamicin. Batches of the growth medium were prepared as needed from a stock solution of modified minimum essential medium. The established cell lines were serially passaged until sufficient material was available for at least four pyrochromatographic experiments.

Pyrolysis of the fibroblast samples was carried out in two stages: first at 400   °C and then at 700   °C. Only the pyrolysis gas chromatograms (PyGCs) from the 700   °C run were used in the study. A typical CF PyGC from a 700   °C run is shown in Figure 2 . The volatile products were separated on a 30 m fused silica capillary column that was temperature programmed.

Figure 2. A pyrochromatogram from a CF heterozygote. Reprinted with permission from Pino, J. A.; McMurry, J. E.; Jurs, P. C.; Lavine, B. K.; Harper, A. M. Application of Pyrolysis/Gas Chromatography/Pattern Recognition to the Detection of Cystic Fibrosis Heterozygotes. Anal. Chem., 1985 , 57 (1), pp 295–302. Copyright (1985) American Chemical Society.

For each subject, triplicate PyGCs were obtained. The 144 PyGCs were standardized (i.e., peak matched) using an interactive computer program. The PyGCs were divided into 12 intervals using approximately 13 evenly spaced peaks that are always present. The retention times of the peaks within each interval were then scaled linearly for best fit with respect to a reference PyGC. The peak matching procedure yielded 214 standardized retention time windows. Therefore, each PyGC was represented as a point in a 214-dimensional space, x  =   (x 1, x 2,   …, x j ,   …, x 214) where x j is the area of the jth GC peak. The CF chromatographic data set – 144 PyGCs of 214 peaks each – was normalized to the percentage of the total peak area and also autoscaled so that each peak had a mean of 0 and a standard deviation of 1 within the entire set of 144 PyGCs.

To apply pattern recognition methods to this data set, the necessary first step was feature selection. For nonparametric linear discriminant functions (i.e., perceptrons, see Chapter 3.18), the classification results will only be significant if the ratio of samples to measurements is greater than 3. Therefore, the number of peaks per PyGC must be reduced to at least one-third of the number of independent PyGCs in the data set. For the final results of the analysis to be meaningful, this feature selection must be done objectively, that is, without using any class membership information.

In this study, it was observed that experimental variables such as cell culture batch number, passage number, donor gender, and column identity contributed to the overall classification process. Other workers have recognized and commented on some of these difficulties in similar situations. 22–24 The confounding of the chemical information by experimental artifacts was investigated through the following experimental sequence.

One way of looking at the confounding used eigenvector projections of the data. The 144 PyGCs were plotted in a two-dimensional map using the two largest principal components derived from the PyGC data. Two examples are shown in Figures 3 and 4 , where clustering related to batch and column identity is evident.

Figure 3. A principal component representation of the pattern space defined by the PyGC peaks of interval 3. The triangles (normal controls) and the squares (CF heterozygotes) enclosed by ellipse 1 represent fibroblast samples analyzed on the same capillary column. The triangles enclosed by ellipse 2 represent normal samples analyzed on another capillary column. (Five different capillary columns were used in the study.) Samples enclosed by ellipses 3, 4, and 5 represent fibroblasts that were grown with different batches of growth medium. Reprinted with permission from Pino, J. A.; McMurry, J. E.; Jurs, P. C.; Lavine, B. K.; Harper, A. M. Application of Pyrolysis/Gas Chromatography/Pattern Recognition to the Detection of Cystic Fibrosis Heterozygotes. Anal. Chem., 1985 , 57 (1), pp 295–302. Copyright (1985) American Chemical Society.

Figure 4. A principal component representation of the pattern space defined by the first 60 PyGC peaks. Squares represent the CF heterozygotes and the inverted triangles represent the presumed normal donors. Samples enclosed by ellipses 1, 2, 3, and 4 represent fibroblast samples analyzed by different capillary GC columns. Five different capillary columns were used in the study. Reprinted with permission from Pino, J. A.; McMurry, J. E.; Jurs, P. C.; Lavine, B. K.; Harper, A. M. Application of Pyrolysis/Gas Chromatography/Pattern Recognition to the Detection of Cystic Fibrosis Heterozygotes. Anal. Chem., 1985 , 57 (1), pp 295–302. Copyright (1985) American Chemical Society.

A second way of studying the confounding involved reordering of the data set according to experimental variables rather than the CF heterozygote versus normal classes. In such experiments, the data set was reordered so that one class contained only those PyGCs analyzed on a particular capillary column and the other class contained all other samples. A discriminant was then developed using PyGC peaks from the reordered data, and the classification success rate was noted. (The data set in the classification problem involving GC columns is referred as reordered data.) Similar reordering experiments were carried out for batch number, passage number, and gender. From these studies, it was learned that several sets of peaks useful for discriminant development in the CF heterozygote versus normal classification problem also supported discriminants that could differentiate PyGCs according to experimental variables. For example, the 12 PyGC peaks comprising interval three of the PyGCs yielded 100% correct classification for CF versus normal as well as the capillary GC column or particular batch of growth medium used to prepare the skin fibroblasts. However, descriptor sets for discriminant development that only yield favorable classification results for the CF heterozygote versus normal classification problem could be found.

Regression analysis using indicator variables for the experimental parameters was also performed to see whether or not there was a significant effect for a particular parameter. For example, a set of PyGC peaks used for discriminant analysis was regressed against an indicator variable constructed for gender. An indicator variable is a variable of zeroes and ones, where the zeroes correspond to one class and the ones correspond to the other. The residuals from this regression analysis were stored and ostensibly only gender information was removed. Discriminants were then developed using these residuals to represent the PyGCs. The classification success rate of a discriminant developed from the residuals was noted and compared to a discriminant developed from the corresponding PyGC peaks themselves. In the case of gender, a 20% reduction in the classification success rate was observed compared to a decision function developed from the corresponding PyGC peaks as descriptors. Such a marked difference in classification power suggests that gender has a strong effect. Regression studies were carried out for passage number as well and similar results were obtained.

Notwithstanding the effects of the experimental variables described above, a discriminant has been developed that differentiates between the PyGCs from CF heterozygotes and that from presumed normal donors essentially on the basis of legitimate chemical differences between the two groups. The development of this discriminant is described below.

All of the PyGC peaks present in at least 90% of the PyGCs were used as the starting point for the analysis. The ability of each PyGC peak alone to discriminate between PyGCs from CF heterozygotes and that from presumed normals was assessed. Data reordering experiments for gender, passage number, and column identity were carried out, and the dichotomization power of each of the 65 GC peaks with respect to these experimental variables was also evaluated. These reordering experiments showed that PyGC peaks were most subject to experimental conditions, and these peaks were not considered for further analysis. The 12 PyGC peaks that produced the best individual classification results when the PyGCs were classified as either CF carriers or normals were used for discriminant development. These 12 peaks spanned the entire PyGC. Variance feature selection 25 combined with the linear learning machine and the adaptive least squares method 26 was used to remove peaks not relevant to the classification problem. A discriminant that only misclassified 8 of the PyGCs (136 correct out of 144 or 94%) was developed using the final set of six PyGC peaks.

The contribution to the overall dichotomization power by experimental parameters of the decision function which was based on just six PyGC peaks was assessed by reordering experiments. In one study, the set of PyGCs was first reordered in terms of donor gender, and poor classification success rates were obtained. Next, the PyGCs were arbitrarily assigned to one of two classes and the discriminatory power of the decision function for this data was then determined. There was little difference in the classification success rate of the decision function for these two sets. That is, reorderings of the PyGCs randomly or by donor gender are essentially equivalent in terms of separability of the data into classes. Similar studies were performed for passage number and column identity and comparable results were obtained. The results of all of these reordering experiments suggest that the decision function based on the six PyGC peaks mainly incorporates chemical information in separating the PyGCs of CF heterozygotes from those of the normals.

The ability of the decision function to classify a simulated unknown sample was testing using segmented cross-validation. Twelve sets of PyGCs were developed by random selection, where the training set contained 44 triplicates and the prediction set contained the remaining 4 triplicates. Any particular triplicate was present in only one prediction set of the 12 generated. Discriminants were developed using the training set data and tested on the prediction sets. The average correct classification for the prediction sets was 87%. This same experiment was repeated again except that members of the prediction set included triplicate samples analyzed on the same column or grown in the same batch of growth media. The average correct classification for the prediction sets in this set of runs was 82%. Although the predictive ability of the decision function was diminished when these confounding effects were taken into account, favorable results were still obtained.

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780444527011000272

Chemical biotechnology • Pharmaceutical biotechnology

Alan J Dickson , in Current Opinion in Biotechnology, 2014

Metabolomic approaches indicate the metabolic changes associated with different phases of growth in batch culture

In simple batch culture, cells undergo a period of exponential growth (perhaps with an initial lag phase) before entering a stationary period (non-growth) before entering a decline (death) phase. Sengupta et al. [25], working with a GS-deficient CHO cell platform making a recombinant antibody (in a proprietary medium), have shown that exponential phase was associated with high rates of glycolytic flux from glucose to lactate with the transition to stationary phase being associated with increased flux (from glucose and 3-carbon intermediates) into the citric acid cycle. Others working with GS-CHO cells have reported a similar switch with production of citric acid cycle intermediates (citrate, fumarate, succinate, malate) especially prevalent during the stationary phase of cell culture [6,28• ]. The simultaneous profiling of intracellular and extracellular metabolites has proven powerful in pointing towards changes to cellular processes, such as changes to intracellular glycerol-3 phosphate in relation to phospholipid biosynthesis/cell growth [6]. As stationary phase has been associated with a more productive phenotype in CHO cells [41•• ], this knowledge sets opportunities to design media that enforce a stationary phase phenotype (or equally to select/engineer cells with a specific metabolic phenotype).

Read full article

URL:

https://www.sciencedirect.com/science/article/pii/S0958166914001062

A Single-use Strategy to Enable Manufacturing of Affordable Biologics

Renaud Jacquemart , ... Jim Stout , in Computational and Structural Biotechnology Journal, 2016

2.1.1 Right-sizing Upstream Processes

For an upstream process using fed-batch cell culture, process scale-up can be achieved by a faster cell culture production rate using multiple bioreactors that are harvested at a certain frequency per week ("run rate"), but this puts strain on the purification process or requires larger or dual column/filtration operations to accommodate the biomass harvest frequency. A more common scale-up method to increase manufacturing capability is via larger working vessels, which requires developing a dedicated bioprocess facility through expansion or new construction [28]. For example, for commercial scale mAb production, the traditional fed-batch bioreactor usually cultures cells in 10,000–25,000   L stainless steel tanks for 7–21   days with a product yield of 2–6   g/L [18]. The renewed interest in perfusion bioreactors together with cell culture advancements such as high titer mammalian cell lines, transgenic expression, and microbial expression can significantly improve the upstream productivity per unit volume and reduce the requirement of high-volume units [29]. Perfusion reactors utilize a continuous supply of cell culture media while the growth-inhibiting by-products are constantly removed over a prolonged production phase (typically >   20   days) to achieve 10–30 times higher cell density compared to a fed-batch reactor [30]. Longer run duration and the ability to sustain biomass levels enable perfusion bioreactors to offer a productivity advantage (in terms of mg/L/d) of at least 4-fold as compared to a fed-batch unit with the same reactor volume [18]; therefore, the same product quantity can be produced with less space and capital cost. The continuous nature of a perfusion operation also makes it a great candidate for continuous processing of biopharmaceutical proteins. Implementation of perfusion reactors has been successfully commercialized, ranging from large biopharmaceutical companies such as Pfizer, Genentech, Shire, and Genzyme/Sanofi [31–33] to small companies and innovative vaccine manufacturers such as CMC Biologics and Crucell [34,35]. Although there are still drawbacks to the technology such as usage of large volumes of medium, and high level of operator training required due to the complexity and intensity of the operation, the economic gain from smaller vessels and facilities has the critical impact on process considerations [18,36].

Read full article

URL:

https://www.sciencedirect.com/science/article/pii/S2001037016300289

Production technologies for monoclonal antibodies and their fragments

Dana C Andersen , Dorothea E Reilly , in Current Opinion in Biotechnology, 2004

Apoptosis can be the major mechanism for cellular death in fed-batch cell culture. Consequently, several different approaches have been recently evaluated to prolong culture viability and, potentially, increase product titers. Members of the Bcl family, specifically bcl-2 and bcl-xL , have represented the most common elements for engineering apoptosis resistance (see [37]), although recent studies have also demonstrated effects using Aven [38], XIAP and CrmA [39] and approaches targeting Caspase-3 at the RNA level have been reported [40]. Another study showed a modest benefit of hsp70 overexpression on NS0 apoptosis-resistance and on the productivity of hybridoma cells generated from these hosts [41]. Following earlier work in which Bcl-2 expression was shown to improve antibody and recombinant protein expression in the presence of butyrate in CHO cultures, an apotosis-resistant, DHFR- and Bcl-2 overexpressing host was recently created by Lee and Lee [42]. In one recent study, the relative merits of Bcl-2 and Bcl-xL were evaluated using a serum-free CHO-DG44 line producing a soluble intercellular adhesion molecule. In this analysis, Bcl-xL was observed to have more potent apoptosis-resistance effects, although benefits on total recombinant protein expression were only observed after amplifying the Bcl-xL expression to relatively high levels [43]. In response to potential concerns about the long-term effects of Bcl-xL overexpression, another group recently developed an inducible system, using the metallothionein promoter, for Bcl-xL expression and apoptosis-resistance during the production phase of hybridoma cultures [44]. Finally, the combination of growth control and anti-apoptosis effects via genetic engineering was recently evaluated in perfusion NS0 cultures using combined p21 and Bcl-2 overexpression and enabled antibody specific productivities approaching 50   pg/cell/day [45•].

Read full article

URL:

https://www.sciencedirect.com/science/article/pii/S0958166904001089

The art of antibody process development

David E. Steinmeyer , Ellen L. McCormick , in Drug Discovery Today, 2008

Schematic of an antibody. Pfizer's developmental compound tremelimumab is an antibody of subclass IgG2 and is composed of four proteins linked by disulfide bonds. Many antibodies are glycosylated, meaning that polysaccharide structures are covalently attached to the proteins at distinct sites. An antibody can have several domains that have activity, including the variable regions on the arms of the heavy and light chains, which bind to antigens, and the Fc domains, which can mediate effector functions. Adapted with permission from Pfizer, Inc.

Read full article

URL:

https://www.sciencedirect.com/science/article/pii/S1359644608001475

Multivariate PAT solutions for biopharmaceutical cultivation: current progress and limitations

Sarah M. Mercier , ... Mathieu Streefland , in Trends in Biotechnology, 2014

Level 1: MVDA on datasets containing classical bioprocess data

Two studies are limited to using basic readily available process variables in relatively simple applications and were therefore classified at level 1 (Table 1). Gunther et al. [43] applied PCA to a pilot-scale industrial fed-batch cell culture process. A similar study was performed by Nucci et al. [44] on a Bacillus production of penicillin G acylase, where MVDA was used to generate on-line multivariate control charts. In these examples, each dataset contained information exclusively related to the process. MVDA models therefore only assessed quality of technical execution of the process, enabling detection of process faults or deviations. This is relevant for process monitoring in a manufacturing setting, for example. However, because no biological responses were captured to measure product quality or even process performance, the relations and interactions between process parameters and process performance or product quality remain unknown. Corrective actions in the case of a deviation are difficult to implement when understanding of these relations is lacking. These studies can therefore be considered as the first, most basic level of MVDA and PAT.

Read full article

URL:

https://www.sciencedirect.com/science/article/pii/S0167779914000584

Sustainability in the biopharmaceutical industry: Seeking a holistic perspective

Fergal Lalor , ... Edmond Byrne , in Biotechnology Advances, 2019

3.3 Continuous downstream processing

While continuous cell culture processing in the form of perfusion cell culture is well-established in the biopharmaceutical manufacturing sphere, continuous downstream purification processes are uncommon, with few documented in literature. While the current paradigm is for large, fed-batch cell cultures feeding similarly large batch purification trains, the previously mentioned economic risks to the industry from biosimilars and unsustainable treatment costs has led to a desire for increased volumetric productivity, reduced capital expenditures and an overall reduction in cost of goods ( Klutz et al., 2015). While single-use technology satisfies this in part, by reducing facility footprint and capital expenditures, it does not impact the issue of volumetric productivity (Klutz et al., 2015). As a result, the adoption of continuous manufacturing becomes attractive. This can be accomplished by employing perfusion cell culture in tandem with continuous purification unit operations such as tangential flow filtration and chromatography. Fully continuous processes have been documented in the literature (Klutz et al., 2015; Klutz et al., 2015; Walther et al., 2015) and continuous chromatography operation also has been documented (Steinebach et al., 2016). Table 5 describes the benefits and challenges of implementing continuous purification.

Table 5. Continuous downstream purification: benefits and challenges.

Benefits Challenges
Increased volumetric productivity Potential economic benefits may not be realised
Increased separation efficiencies Regulatory challenges from product licence changes
Increased product quality Development of reliable methods for all unit operations
Reduced capital expenditures Increased complexity
Debottlenecking the process

Continuous purification processing is plausible utilising current technologies (Jungbauer, 2013). Significant challenges exist to its implementation at a commercial scale, ranging from technical issues including the development of a reliable method of incubation for steps such as viral inactivation and diafiltration (Przybycien and Titchener-Hooker, 2015) to uncertainties surrounding the predicted economic benefits being realised at commercial scale (Jungbauer, 2013) and, perhaps most significantly, the regulatory requirements to change product registrations (Przybycien and Titchener-Hooker, 2015). Until commercial scale applications of continuous purification processing are implemented and analysed, its development and uptake may remain inhibited by established attitudes within the industry.

Read full article

URL:

https://www.sciencedirect.com/science/article/pii/S0734975019300515