Combination of a bioreactor with a flow microcalorimeter

Cinderalla Solution

Natural Metabolism Boosters

Get Instant Access

By 1997, Kemp et al [19] had developed a solution for measuring the heat flow rate of animal cells growing under the controlled conditions of the bioreactor that is independent of the size of the culture vessel. It allows the measurement of the heat produced in industrial-scale plant by circulating the cell suspension from the bioreactor to the flow calorimeter and then returning it. As a model for a large-scale facility, a 3-L Applikon bioreactor containing a stirred suspension of Chinese Hamster Ovary cells (CH0320) genetically engineered to produce interferon-y (IFN-y) was connected by thermostated PEEK transmission tubing (see Fig. 1 [12]) to an ex situ, on-line Thermometric Thermal Activity Monitor (TAM) heat conduction, twin differential, flow microcalorimeter [31] (Thermo-

Bioreactor Agitator Dimensions

Fig. 1A schematic diagram for the measurement of the heat flow rate of bioreactor-cul-tured animal cells using a flow microcalorimeter. 1 - bioreactor; 2 - cultured cells; 3 - jacket waier for temperature control in the bioreactor; 4 - agitator; 5 - the outlet tubing of the jacket water is used for warming the cell suspension in the tubing leading to the microcalorimeter; 6 - a non-conductive sponge-plastic pipe is used to reduce heat dissipation along the outlet jacket water attached to the PEEK tubing transmitting the cell suspension to the calorimeter; 7 - PEEK T-piece; 8 - PEEK two-way valve; 9 - PEEK tubing (1 mm I.D.); 10 - glass bottle holding sterilized medium for washing the flow vessel of the microcalorimeter through short-time interruption of the heat flow measurement; 11 - sterile medium for cleaning the flow vessel free of possible accumulated animal cells, that would result in an overestimation of the heat flow rate for the cells in the bioreactor; 12 - a representation of the 4-channel Thermometric calorimeter (TAM); 13 - the gold flow vessel assembly; 14 - the dotted boundary, enclosing essentially the bioreactor and the flow vessel of the microcalorimeter, is an open thermodynamic system for enthalpy balance studies; 15 - the peristaltic pump (Reproduced from [12] with permission)

metric AB, Jarfalla, Sweden) with a gold flow-through vessel that had been modified for downward flow to try and minimise blockage by cellular debris. The early experiments were performed with the commercially available vessel that has a nominal capacity of 0.6 cm3 but a customised one of larger volume became available for later work. It is not always understood that the thermal volume of a calorimetric vessel may differ from the physical one for various reasons. Wadso has frequently emphasised this point (for instance, in [32]) and advocated that, in addition to the daily electrical test, chemical calibration should be undertaken for a newly delivered vessel under the exact experimental conditions and repeated whenever there was an alteration in these conditions. He found that the slow imidazole-catalysed hydrolysis of triacetin is a suitable reaction [32]. Using this reaction, Guan et al. [33] found that the thermal volume of the standard gold flow-through vessel at a pumping rate of 35 cm3 h-1 was 0.77 cm3, nearly 30% greater than the nominal volume. More recently, O'Neill et al. [34] showed that the kinetics of the triacetin hydrolysis reaction is too slow for the calibration of flow vessels with their relatively short residence time and suggested the use of the base catalysed hydrolysis of methyl paraben with faster kinetics as an alternative. In a subsequent paper, O'Neill et al. [35] showed that the substance that filled the reference vessel of the twin differential micro -calorimeter affected the thermal volume of the vessel, often in combination with the pumping rate. Indeed, these authors maintain it could be the only factor if the reference vessel was filled with water or medium rather than air, and the range of flow rates was restricted to below 50 cm3 h-1. This implies that, at the 35 cm3 h-1 pumping rate employed by Guan et al. [12, 33], the thermal volume equalled the physical volume. Occasionally, the flow vessel in their hands suffered from blockages that were probably due to the narrow internal diameter (ID = 1 mm) compounded by the comparatively slow pumping rate.

In more recent studies in Aberystwyth, a specially fabricated flow module kindly provided by Thermometric AB to Kemp's design (see Fig. 2) has been used as the standard [36]. Except for the peristaltic pump itself that has Viton tubing, the flow line was 1.5 mm (id) stainless steel throughout and was optimised for a pumping rate of 100 cm3 h-1, a speed that was achieved without undue noise in the signal. At this rate, chemical calibration with the tiracetin reaction gave an effective (thermal) volume of 1.05 cm3, very close to the nominal volume of 1.00 cm3 [36] Most recently, the vessel has been re-calibrated with the methyl paraben hydrolysis reaction and found to be within the specified vol-

Thermal Activity Monitor 2277
Fig. 2(A) A pictorial representation of the new flow module for the Thermometric 2277 Thermal Activity Monitor. The section labeled by A-A is illustrated in Fig. 2(B)

Peltier ceil

Peltier ceil

Measuring side Reference side

Fig. 2(B) A sectional view of the layout of the thermal detector in the new flow module for Thermometric 2277 TAM. The vertical position of this section is labeled in Fig. 2(A) as A-A. (Reproduced from [36] with permission)

Measuring side Reference side

Fig. 2(B) A sectional view of the layout of the thermal detector in the new flow module for Thermometric 2277 TAM. The vertical position of this section is labeled in Fig. 2(A) as A-A. (Reproduced from [36] with permission)

ume [34]. Although Chisti [37] drew attention to the dangers of hydrodynamic damage to cells in such situations particularly with respect apoptosis, but also purely mechanical damage, exhaustive tests of prolonged exposure of the cells to pumping over several hours through the tubing failed to reveal any such deleterious changes [33, 36].

When the heat flow rate data are obtained simultaneously with the measurements of the concentrations of mat erials (substrates and products) in the bioreactor, the validity of the results can be tested by calculations described as the enthalpy balance method (see Section 5; also [26, 38]). Bearing in mind that the system is open in the thermodynamic sense [39], it is important for the validity of these calculations to (i) define the boundary of the system and (ii) ensure that the conditions are the same throughout that system. For the first requirement (see Fig. 1), it was decided that the infinitely thin boundary was in the walls of the glass bioreactor and the tubing. The nature of the latter depends on the flow system. For the initial studies, it was (in sequence) the PEEK transmission tubing, the Viton (or Aliprene) of the pump tubing and the gold in the wall of the flow vessel. For the customised vessel, the boundary was in the stainless steel and the Viton tubing. To help maintain similar conditions throughout the system, the transmission tubing was thermostated at the same temperature as the bioreactor and the calo rimeter, while the materials for the transmission lines and the pump tubing were selected mainly for their low gaseous diffusivity. In the initial studies, Guan et al. [33] found experimentally that pumping the cell suspension at a rate of 35 cm3 h-1 was sufficient to ensure that the cells did not consume so much of the dissolved oxygen in the medium to drive the concentration of it below normoxic levels before the cells were returned to the bioreactor. Even so, a higher rate could only be beneficial to the cell metabolism, providing that there was no effect on viability caused by physical trauma [37]. In trials with the customised vessel, this proved to be the case [36].

Kemp and Guan [40] recorded the continuous trace of the heat flow rate and compared it with the numbers of viable cells measured at discrete time intervals during the batch culture of recombinant CHO 320 cells. The important conclusion from these assessments was that the heat flow rate as the indicator of the metabolic rate declined while there were still increases in the cell number concentration (see Fig. 3). The implications were that the heat flow rate could be an even more sensitive reflection of cellular metabolism if it were made specific to biomass, i.e. scalar heat flux.

Microcalorimeter

Fig. 3 The heat flow rate of growing cells measured on-line by the microcalorimeter and scaled to the unit bulk volume of RPMI 1640-based culture medium buffered with 20 mM HEPES and 4 mM bicarbonate (—). Estimates were made for the number of viable cells per cm3 bulk volume (o ) at discrete time intervals (Reproduced from [40] with permission)

Fig. 3 The heat flow rate of growing cells measured on-line by the microcalorimeter and scaled to the unit bulk volume of RPMI 1640-based culture medium buffered with 20 mM HEPES and 4 mM bicarbonate (—). Estimates were made for the number of viable cells per cm3 bulk volume (o ) at discrete time intervals (Reproduced from [40] with permission)

Counting cell numbers by haemocytometry is a labour-intensive, time-consuming and a relatively inaccurate (± 10%) off-line method to assess biomass. Over the years, the Institute of Biological Sciences in Aberystwyth has been interested in the biological applications of dielectrics [41] and, after a series of trials, it was discovered that the animal cell biomass could be measured satisfactorily with the in situ, on-line probe of a radio frequency (0.5 MHz) dielectric spectrometer. Aber Instruments Ltd. (Aberystwyth, Wales, UK) then marketed it as the Viable Cell Monitor. The capacitance (C, in Farads, F) measurements made with this instrument have been shown both theoretically [41] and by the essential calibration of particle size with a flow cytometer [12]to record the volume fraction of the viable cells, providing there is no change in conductivity and also no alteration to the cell volume during the culture period. Then, monitoring the change in capacitance indicates the variation in biomass.

In the experiments, the signal from the spectrometer probe in the bioreactor was digitised and fed to the Applikon BioXpert software used as a log for the culture monitoring data [12]. The capacitance signal was blanked for the value obtained with the growth medium, both this signal (AC) and the one from the calorimeter were smoothed by the moving average technique in the software and heat flow rate was divided by the digitised capacitance signal (biomass) to give the flux. The resulting trace, J®/C, was comparable with the control in which the heat flow rate at discrete times during the culture was divided by the cell number concentration (N) obtained off-line with a Coulter counter, J®/N, (Fig. 4) and it had the advantage of being continuous, on-line [12, 23]. The on-line heat flux signal was then compared after the experiment with the discrete values for the changes in the concentrations of the major substrates, glucose and glutamine, measured by off-line methods (Fig. 5). It was highly significant that the heat flux monitor detected a decline in the metabolism of the CHO 320 cells before there was an apparent alteration in the consumption of the two substrates. This result confirmed the fact that, despite evidence of reduced metabolic flux, the cells continued to grow (see Fig. 3) and produce IFN-y.

It should be recognised that others have also adopted dielectric spectrometry to measure animal cell biomass on-line but so far have not combined it with heat flux measurements. For instance, Ducommun et al. [42] indirectly determined the concentration of free suspended and immobilised CHO SSF cells by this on-line method and also applied it successfully to two industrial high density culture pro-

Fig. 4 On-line heat flux measurements adjusted to per cm3 bulk volume (—) and heat flow per viable CHO 320 cell (o) over 140 h of a batch culture. In the expression of J0/C, the values are given in terms of 1 cm3 bulk medium volume (Reproduced from [12] with permission)

0 20 40 60 SO 100 120 140

111'K ill

Fig. 5 Comparison of the heat flux (Jorn) with the fluxes of glucose (Jolc), glutamine (Join) and IFN-y (JIFN), as well as the specific growth rate (|J.) during the batch cultivation of CHO 320 cells in suspension. Heat flux (o), glucose flux (□), glutamine flux (A), IFN-y flux (x) and specific growth rate (•). The bars indicate the period over which the discrete off-line measurements were made to give the individual average values for fluxes (Reproduced from [12] with permission)

0 20 40 60 SO 100 120 140

111'K ill

Fig. 5 Comparison of the heat flux (Jorn) with the fluxes of glucose (Jolc), glutamine (Join) and IFN-y (JIFN), as well as the specific growth rate (|J.) during the batch cultivation of CHO 320 cells in suspension. Heat flux (o), glucose flux (□), glutamine flux (A), IFN-y flux (x) and specific growth rate (•). The bars indicate the period over which the discrete off-line measurements were made to give the individual average values for fluxes (Reproduced from [12] with permission)

cesses, using it to determine on-line the concentration of CHO cells immobilised on macroporous microcarriers in a stirred bioreactor and in a packed-bed of immobilised hybridoma cells on disk carriers [43]. For the latter, dielectric spectrometry was used as a tool to characterise the packed-bed process, showing for instance the maximum cell concentration that could be reached was 2.0x10n cells per kg of disk carrier. From the future perspective of combining dielectric spectrometry with calo-rimetry, the more exciting finding from von Stockar's laboratory [44] was that the instrument could be modified to scan CHO perfusion cultures every 20 min over the excitation frequency range of 0.2 MHz to 10.0 MHZ to give both maximum and zero cell viability and signify the end of lactate consumption. The combination of scanning dielectric spectrometry (SDS) and calorimetry could give profound insight into the relation between the metabolic rate of the cells and their anabolic processes, particularly in the production of heterologous proteins.

Even with the putative power of SDS measurements combined with heat flux, it is often important to know the changes of crucial metabolites, such as glucose, glutamine and lactate as vital components in the overall process. Guan et al. [12] made discrete measurements of these materials but, of course, the profile of the concentration changes is of limited value compared with the fluxes. In order to reveal changes in the rate of change in material, normally the values for the concentrations of a given substance must be differentiated at two time points, t1 and t2 for each period of the cell culture. Although this procedure will give the rate, it is only the average, relative to that time period, as opposed to the instantaneous rate available from the calo rimeter. Data for the material concentrations can be converted to the instantaneous rate by using the Tikhonov-Phillips integrative method of smoothing [45, 46].

Guan et al. [12] took the estimations for the changes in the catabolic substrates and products during the growth of CHO 320 cells, obtained the average relative rates and made them specific to cell number concentration. The calculations (see Fig. 5 and [23]) confirmed the impression that the specific growth rate (|) and the fluxes of the major substrates and products, including IFN-y began to decrease at the relatively early time in the culture indicated by the heat flux biosensor. The exact mathematical relationship of the heat flux to the material fluxes was established by plotting the former fluxes against the collection of the latter. It was then clearly seen from Fig. 6 that heat flux is a function of the material fluxes in a monotonic relationship. From this, Kemp and Guan [25] concluded that heat flux can be utilised in cell culture as an on-line, real time probe which: (i) has a rigorous thermodynamic foundation to it; (ii) gives the instantaneous metabolic flux; (iii) is non-invasive as well as being non-destructive; and (iv) is robust with no consumable components. Therefore, it can be used to monitor cultures and improve the design of media as well as being made an effective control variable for fed-batch cultures.

Fig. 6 Heat flux is plotted as a function of the specific growth rate (■) showtng the monotonically decreasing relationship. This dependence extends to the fluxes for 104 x IFN-y production flux, IU s-1 per cell, (a) and the major substrates, 107 x glucose consumption flux, mol s-1 per cell, (♦) and 107 x glutamine consumption flux, mol s-1 per cell (a) (Reproduced from [23] with permission)

Fig. 6 Heat flux is plotted as a function of the specific growth rate (■) showtng the monotonically decreasing relationship. This dependence extends to the fluxes for 104 x IFN-y production flux, IU s-1 per cell, (a) and the major substrates, 107 x glucose consumption flux, mol s-1 per cell, (♦) and 107 x glutamine consumption flux, mol s-1 per cell (a) (Reproduced from [23] with permission)

Was this article helpful?

0 0
Tips and Tricks For Boosting Your Metabolism

Tips and Tricks For Boosting Your Metabolism

So maybe instead of being a pencil-neck dweeb, youre a bit of a fatty. Well, thats no problem either. Because this bonus will show you exactly how to burn that fat off AS you put on muscle. By boosting your metabolism and working out the way you normally do, you will get rid of all that chub and gain the hard, rippled muscles youve been dreaming of.

Get My Free Ebook


Post a comment