journal of biotechnology and bioengineering volume 1 ... · reduced tanon eps 300 electrophoresis...
TRANSCRIPT
Nessa Publishers| www.nessapublishers.com Page 1
Journal of Biotechnology and Bioengineering
Volume 1| Issue 1
Research Article Open Access
Performance of the AbSolute® High Cap resin in the Nimotuzumab capture step
Martinez Rosario 1*, Ruland Yvan 2, Zhao Gang3, and Shen Chunjuan4 1The Center for Molecular Immunology, Cuba. 2, 3, 4, NOVASEP Asia- Biopharma Business Unit, Shanghai, 201203 China. *Corresponding author: Martinez Rosario, The Center for Molecular Immunology, Cuba.
Email: [email protected]
Citation: Martinez Rosario (2017) Performance of the AbSolute® High Cap resin in the Nimotuzumab capture
step: Nessa Journal Biotechnology and Bioengineering.
Received: February 10th 2017, Accepted: March 7th 2017, Published: May 22th 2017
Copyright: © 2017 Martinez Rosario, Yvan Ruland, Zhao Gang and Shen Chunjuan et al. This is an open-
access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are
credited.
Abstract
Protein A is the affinity chromatography ligand of choice for first-step capture in the purification of mAbs as its
high selectivity gives excellent purity with high yields. Furthermore,a Protein A capture acts as a key volume
reduction step in the process since it concentrates significantly the product stream. The relatively high cost of
suchaffinity resins leads to consider different operational strategiesin order to well optimize this step. One of
them is to evaluate Protein A matrixes with highperformance in terms of dynamic binding capacity (DBC) at
high flow rates.
Recently a new Protein A ligand, AbSolute® High Cap (AbSolute HC), was introduced by Novasep Company,
which is modified with respect to high binding capacity at higher velocities. This article describes the
performance of AbSolute HC gelfor Nimotuzumab capture, by measuring its Dynamic Binding Capacity (DBC)
at different flow rates and feed concentrations at lab scale. AbSolute HC data are also compared with
MabSelectSuRe, the affinity gel used in the current purification process. The process conditionsare adjusted
including the maximum speed for all steps with an assessment of impact on Nimotuzumab purity using AbSolute
HC. In addition the impact of loading amount on aggregates formation with this media and the resin lifetime are
evaluated. The results show a better DBC at 10% on AbSolute HC than MabSelect Sure, with a higher DBC at
10% on Absolute HC with IgG concentration at 2.3 g/L than 0.13 g/L. No significant loss of purity or yield for
speeds until 1000 cm/h for all steps is observedusing AbSolute HC resin. The lifetime of the media is tested up to
200 cycles with suitable results using NaOH and PAB solutions as CIP buffers.
Key words: Protein A, mAbs, affinity chromatography, AbSolute HC, dynamicbinding capacity (DBC), lineal
flow rate, yield, aggregates.
Nessa Publishers| www.nessapublishers.com Page 2
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
Introduction
The capture step in the downstream processing of monoclonal antibodies (mAbs) is often the bottleneck and the
most expensive step due to the use of Protein A media. Selection of a most suitable affinity resin based on
binding capacity and affinity is typically performed prior to optimization.
In recent years, several new-generation Protein A media have been launched on the market claiming improved
mAb capture, or improved resistance to cleaning agents. AbSolute® High Cap, the new Protein A media by
Novasep, developed for the capture of monoclonal and polyclonal antibodies from large volume fermentation
feedstock, maintains excellent DBCs at high flow rates, therefore providing substantially improved productivity
and strongly reduced costs compared with all existing protein A media.
AbSolute HC is based on a modified silica matrix that exhibits strong mechanical, thermal and chemical
stability. The particles are small (35 μm) with a perfect spherical shape and very narrow size distribution. Pore
size is 100 nm and shows narrow size distribution (± 15 nm). Optimized particles and pores allow high efficiency
and faster mass transfer which increase the Dynamic Binding Capacity (DBC) (1).
Particles Cross Section Surface of Particle
Figure 1: Pictures of AbSolute®High Cap beads
The matrix is rigid and incompressible so a 1L column is packed with 1L of AbSolute HC and the media does
not shrink or swell in different solutions. These exceptional mechanical properties combined with a unique
kinetic performance allow operations at higher velocities (up to 1,500 cm/h, depending on the column size and
equipment).
Its resistance to high pH levels makes it stable through alkaline cleaning and sanitization. Indeed, AbSolute HC
remains stable after 400 cycles of use with alkali washing using 100 mM NaOH + 0.5 M NaCl every 10 cycles(2).
Also Novasep reports that using AbSolute HC gel implies an optimization of the geometry of all columns, thus
minimizing the volume of protein A required to perform the step, in order to maximize productivity. In these
conditions, performing the step at industrial scale using AbSolute® High Cap allows a volume reduction in the
media and an increase in productivity by up to 2.7-fold compared with other industry standards, resulting in a
cost savings of up to $1,000,000 for the first load of media(2,3).
Nessa Publishers| www.nessapublishers.com Page 3
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
This study outlines the performance of AbSolute HC resin for Nimotuzumab capture by measuring its DBC at
different flow rates and feed concentrations at lab scale. AbSolute HC data are also compared with Mab Select
SuRe, current affinity gel used in the purification process.
The impact of mobile phase velocity and of loading amount on product quality is evaluated as well, with the
product critical quality attributes (CQA’s) being measured by yield, host cell protein (HCP) clearance, and
percent of aggregate.
This report also evaluates the lifetime of the media, tested up to 200 cycles using NaOH and PAB solution
(Phosphoric acid, Acetic acid, Benzyl alcohol mixture). Lifetime is assessed by measuring the variation of DBC
over the time, but product quality in terms of purity is monitored as well over the whole lifetime study.
Materials and Methods
The supernatant is obtained from a perfusion bioreactor of 1000 L of mammalian cell culture. The harvests
filtrated are tested at 0.13 g/L IgG concentration, and at 2.3 g/L IgG concentration after TFF concentration.
Absolute HC and Mab Select Sure resins are packed into columns of 0.5 cm of internal diameter (i.d). Column
height is 10 cm for all experiments except for lifetime study which is carried out with a 5 cm height column.
The equipments used for this study, performed in NovaSep facilities in Shanghai, are listed in the Table 1:
Table 1: List of Equipments used
Equipment Supplier
AKTA Purifier GE Healthcare
Analytical HPLC Waters
Tricorn 5 mm ID columns (5 and 10 cm height) GE Healthcare
Balances Sartorius
SDS Page imaging system Tanon-2005 Tian Neng
Microplate reader Spectramax 190 MD
pH meter/conductimeter Mettler Toledo
Membrane TangenX Novasep
TFF skid Novasep
Buffers used at the different steps are shown in Table 2:
Nessa Publishers| www.nessapublishers.com Page 4
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
Table 2: Buffers and step in the experimentation
Step Buffer
Loading Nimotuzumab harvests filtrated
Equilibrium 50 mM Tris 150 mM NaCl, pH 7.3-7.5, Cond. 15-20 mS/cm
Washing 1 50 mM Tris 1 M NaCl, EDTA 10mM, pH 7.3-7.5, Cond. 80 - 90
mS/cm
Washing 2 50 mM Tris, 150 mM NaCl, pH 7.3-7.5, Cond 15-20 mS/cm
Elution 0.1 M Citric acid, pH 3.2~3.8, Cond 3-7 mS/cm
Wash Purified water
Cleaning NaOH 100 mM
Storage 20 % ethanol
Experimental Conditions
Analytical Assays:
Table 3 summarizes the analytical assays, samples, methods and equipment used in the experimentation.
Nessa Publishers| www.nessapublishers.com Page 5
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
Table 3: Analytical assays evaluated
Analytical Sample Method Equipment/column/reagent
items
Feed,
HPLC-RP
Column: Zorbax SB 300 C8,
4.6*250mm,5µm
Flow through/wash
HPLC: waters alliance 2695-996
IgG
for BTC test
Solvents: TFA/ACN; Gradient
concentration
elution
Eluted IgG
UV280
Spectrophotometer
Supplier: APL instrument
Reduced Tanon EPS 300 electrophoresis
SDS- system
Purity Eluted IgG
PAGE Tanon 2500 gel Image system
Column: TSK G3000SWxl
SEC 7.8*300mm
HPLC: Waters Alliance 2695-996
ELISA Kit: Cygnus F220.
HCP Eluted IgG ELISA Equipment: Molecular Devices ,
Spectra Max 190
Protein A ELISA Kit: Cygnus F400
leakage Eluted IgG ELISA Equipment: Molecular Devices ,
Spectra Max 190
All the samples were filtered by 0.2 µm on sterile bottle and stored at 4°C.
Nessa Publishers| www.nessapublishers.com Page 6
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
Breakthrough curves (BTC’s) at different flow rate – Experimental conditions:
The buffers and steps used for those experiments are listed in table 2 above.
A characterization of the capture step of Nimotuzumab is performed at lab scale from a cell culture supernatant,
by tracing breakthrough curves under different conditions of loading linear velocities.
Feed is loaded continuously onto the column, collections are made at regular interval of time at the column
outlet. The concentration in the breakthrough C is measured by HPLC-RP (see table 3 above) and the dynamic
binding capacities Q (x-axis, mg/mL gel) are calculated using the formula:
Q = C0 * t * F / CV Where:
Q: Dynamic binding capacity (mg/mL resin).
C0: Initial protein concentration (mg/mL).
t: Time (min).
F: Volumetric flow rate (mL/min).
CV: Column volume (mL).
When the concentration C in the breakthrough is approaching the initial concentration C0, the column is fully
saturated. It is washed until the UV baseline is reached and then eluted. A regeneration and equilibration is
performed before the next BTC.
Beside the concentration C in the breakthrough, the concentration is also measured in the Wash stream and
Elution stream.
Impact of speed on purity on AbSolute HC – Experimental conditions:
The buffers and steps used for those experiments are listed in table 2 above. A total of 10 experiments are
performed, by varying the mobile phase velocity between 300 and 1000 cm/h. In some experiments the velocity
is kept the same for all steps (load, wash, elution, CIP), in others the velocity is increased for the wash or elution
step.
The time for loading step is adjusted so that 60 to 80 % of the total capacity previously determined in BTC’s
experiments is reached.
After elution, the product is incubated during 1 hour at low pH condition (elution buffer at pH 3.2-3.8) at room
temperature for viral inactivation. Then a microfiltration using 0.22 µm filter is performed to reduce bioburden
content.
Nessa Publishers| www.nessapublishers.com Page 7
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
Before SEC, SDS Page, HCP and Protein A analysis (see table 3), the samples collected at elution step are
desalted using following conditions:
Column: Hitrap 5 mL; Resin: Sephadex G 25; Buffer: 20 mM Tris, pH 7.6-8.0, conductivity 0.9-1.2 ms/cm;
Flow rate: 100-150 cm/h; Loading volume: 20-30 % column volume.
Lifetime studies – Experimental conditions:
The lifetime study is carried out on a Tricorncolumn of dimension 5 x 50 mm (GE Healthcare) packed with 1 ml
of AbSolute HC.
200 cycles are performed using the buffer conditions listed in table 2 except the cleaning in place (CIP)
conditions, a mixture of acids and alcohols “PAB” being used instead of NaOH. PAB is a mixture of 120mM
H3PO4, 167 mM acetic acid and 2.2 % benzyl alcohol, applied on the column for cleaning at each cycle with 20
minutes contact time. Every 10 cycles, NaOH CIP is performed as well using a 0.1 mM NaOH solution with 20
minutes contact time as well. When NaOH CIP occurs, the column is washed in between the PAB and the NaOH
cleaning steps. All conditions are summarized in table 4 below.
Table 4: Gel Lifetime conditions used
Step Solution Flow pH Conductivity CV
rate (mS/cm)
PAB PAB: 45 cm/h ~1.5 - 3
Hygienization
[a]
120 mM H3PO4,
167 mM acetic acid,
2.2% benzyl alcohol
NaOH CIP[b] 50 mM Tris-HCl, 150 mM 700 7.3- 15-20 20
/Wash NaCl cm/h 7.5
NaOH CIP[a] 0.1M NaOH, 1M NaCl 60 cm/h 12 - 4
/CIP
[a] PAB hygienization is done for all the 200 cycles. Every 10 cycles, a NaOH CIP is performed as well. Contact
time is 20 minutes in both cases.
[b]Column is washed in between PAB and NaOH CIP.
Nessa Publishers| www.nessapublishers.com Page 8
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
A series of test is performed to check the column performance over the time, as summarized in table 5 below:
Every 20 cycles starting from cycle 1, a BTC is performed under conditions described previously in this section,
in order to measure the variation of 10 % DBC value along the 200 cycles.
The elution stream is analyzed for Nimotuzumab purity every 20 cycles starting from cycle 20.
The product eluted at cycles 31, 33, 35, 37 and 39 is also analyzed, in order to evaluate the impact on quality of
Nimotuzumab when only PAB solution is used as CIP test, and no NaOH.
Table 5: tests performed during lifetime study with identification of corresponding cycle number
Results and Discussion
BTC test at different flow rate on MabSelect Sure and AbSolute HC.
The quantity of antibodies that can be loaded per mL of gel is determined by the breakthrough curves performed
at the linear velocities that maximize the overall process productivity. The breakthrough curves of the resins
studied are presented in figures 2 and 3 at the linear flow rate defined and feed concentration C0 of 0.13 and 2.3
g/L IgG.
Nessa Publishers| www.nessapublishers.com Page 9
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
Figure 2: BTC of MabSelect Sure resin at 0.13 and 2.3 g/L IgG in feed.
Figure 3: BTC of AbSolute HC resin at 0.13 and 2.3 g/L IgG in feed.
In all cases, increasing the linear flow rate decreases the dynamic binding capacity. This is mainly due to the fact
that using a faster flow rate increases the resistance to the mass transfer of the protein to the interior adsorption
sites of the matrix beads. Antibodies are very large proteins of low diffusivity in the mobile phase and at higher
speed, a larger number of them will not be able to enter the pores in the matrix beads. As a result of a low
diffusivity, it will also take a longer time for the proteins inside the pores to get out of the bead and undergo
further elution along the column. The higher is the mobile phase speed, the shallower is the adsorption front on
the breakthrough curve and the higher is the concentration of antibody in the breakthrough for a given
loading(4,5).
Numerous chromatographic models have been well described in the literature, to characterize the adsorption
equilibrium and adsorption kinetics (6,7). In the case of AbSolute adsorbent, a modeling approach based on
integration of process modeling and experimentation has been applied and has shown the high capacity
properties of such Protein A adsorbent based on matrix silica (8). As a matter of fact it is remarkable in the figure
3 that the breakthrough curves in AbSolute HC gel are steep whatever the linear velocity, due to a very uniform
particle size distribution and a homogeneous pore size distribution for a better mass transfer.
Nessa Publishers| www.nessapublishers.com Page 10
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
On the same figure we observe on the BTC’s achieved with MabSelect SuRe resin some difficulties to reach the
full saturation. This has been reported for many other Protein A adsorbents(9) with some suggestions related to
molecular stretching and changes in transport mechanism upon adsorption, broad particle size distribution,
competition between monomers and aggregates. The results as shown on figure 3 could support the explanations
related to particle/pore size distribution and competition between monomers and aggregates. As a matter of fact,
a complete saturation is easier to reach with AbSolute HC material of narrow particles and pores distribution,
and seems to be easier to reach at lower concentration.
Table 6 shows the summary of those results at C/C0 0.1 (Q=10 %).
Table 6: Summary result of dynamic binding capacity (Q) for the resins studied.
MabSelect Sure resin Absolute HC resin
IgG conc. Flow rate Q (g/ L resin) IgG conc. Flow rate Q (g/ L resin)
(g/L) in feed (cm/h) at 10 % C/C0 (g/L) in feed (cm/h) at 10 % C/C0
300 26.8 300 42.2
0.13 600 15.5 0.13 600 36.8
1000 10.7 1000 34.1
300 29.3 300 55.0
2.3 600 14.3 2.3 600 50.3
1000 9.9 1000 42.0
As can be seen better results of dynamic binding capacity are obtained with AbSolute HC gel. The highest values
are observed on Absolute HC with IgG at 300, 600 and 1000 cm/h with a range of 42.2-55.0; 36.8-50.3 and 34.1-
42 g/L depending on IgG concentration in feed at concentration at 0.13 g/L and 2.3 g/L.
Nessa Publishers| www.nessapublishers.com Page 11
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
Impact of speed on purity on AbSolute HC
Beside the dynamic capacity at high flow rate, the ability of a Protein A affinity media to increase an antibody
purity is obviously of paramount importance in such media performance evaluation.
Table7, figures 4 and 5 show the results achieved in this phase to evaluate the impact of high velocities on
Nimotuzumab purity and yield using different concentrations in the feed. The loading being the longest step in
the process, increasing this step’s speed will have a dramatic impact on the productivity. Three different linear
velocities are tested for loading: 300 cm/h, 600 cm/h and 1000 cm/h.
In some experiments, a higher speed is applied for either the wash step (Exp #4 and #9) or the elution step
(Exp#5 and #10).As a matter of fact, increasing those steps speed will allow to decrease further the cycle time
with some direct impact on production schedules if the method is implemented at manufacturing scale.
Table 7: Summary table of speed impact on purity with Absolute HC
Exp. Flow rate IgG Yield (%) Purity by Purity by
No. (cm/h) conc. HPLC- SDS-
in feed SEC (%) PAGE (%)
(g/L)
1 300 99,23 97,05 94,85
2 600 98,49 97,53 95,1
3 1000 99,06 96,22 95,18
4
300 cm/h: load, elution and CIP;
600 cm/h: wash
0,13
99,08 97,35 96,85
5
300 cm/h: load, wash and CIP; 600
cm/h elution 99,93 97,33 96,28
6 300
2,3
97,91 97,25 96,56
7 600 98,77 97,66 96,54
8 1000 98,93 97,9 96,59
9
300 cm/h: load, elution and CIP;
600 cm/h: wash 96,78 97,63 95,01
10
300 cm/h: load, wash and CIP; 600
cm/h: elution. 98,6 97,7 95,14
Nessa Publishers| www.nessapublishers.com Page 12
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
Figure 4: Pictures of reduced (right) and non-reduced SDS-PAGE (left) on some experiments.
Figure 5: Chromatogram profile and purity result by HPLC-SEC.
Figure 5: Chromatogram profile and purity result by HPLC-SEC.
As reported in table 5 and figure 4, high purity and recovery values are achieved for each condition evaluated
using Absolute HC resin at both 0.13 and 2.3 g/L IgG concentration on feed (purity being measured by HPLC-
SEC and SDS-Page electrophoresis). No significant loss of purity and yield are observed for speeds until 1000
cm/h for loading, and until 600 cm/h for wash and elution steps (the maximum linear velocity tested for those
steps).
Process-related impurities such as PA leakage and host cell protein (HCP) were also determined in these
experiments. Figures 6 and 7 show the results obtained.
Name Retention
Time
Area %
Area
Height
1 Dimer 6.699 157719 2.14 6963
2 mAb 8.084 7214232 97.70 365039
3 ------ 11.542 12119 0.16 946
Nessa Publishers| www.nessapublishers.com Page 13
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
Figure 6: PA leakage results at different conditions tested with Absolute HC. For each condition, the left column
shows the result achieved at 0.13 g/L feed concentration, and the right column at 2.3 g/L.
Figure 7: HCP results at different conditions tested with Absolute HC. For each condition, the left column shows
the result achieved at 0.13 g/L feed concentration, and the right column at 2.3 g/L.
According to figures 6 and 7 the Protein A leakage is similar for all the tests performed with different velocities.
There is no significant difference either in HCP values for eluted IgG under various operating conditions tested
with the same concentration of IgG in feed solution.
Some slight difference is observed for HCP content in the tests performed with higher feed concentration (2 g/L)
:HCP is increased when a higher elution speed is used than loading and washing. However such HCP content in
the elution stream is still within specifications typically defined for a Protein A capture step.
Nessa Publishers| www.nessapublishers.com Page 14
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
Lifetime study
Cleaning, sanitization, and storage procedures that preserve resin function and integrity are essential for
obtaining reasonable resin lifetimes. The behavior of the DBC and purity are evaluated on 200 cycles. Figures 8
shows the results.
Figure 8: DBC on Absolute HC over 200 cycles.
Factors that influence resin lifetime include the nature of the feed stream and raw materials used in its
purification, the resistance of the resin to cleaning and sanitization agents, and proper storage conditions. During
this test, less than 10 % decrease of 10 % BDC is measured after 200 cycles with 0.1M NaOH CIP solution every
10 cycles for 20 minutes contact time per cycle. Beside NaOH CIP every 10 cycles, a PAB hygienization step is
applied as CIP at each cycle, with also 20 minutes as contact time.
Purity and process-related impurities are also analyzed. Figures 9, 10 and 11 show the results.
Figure 9: Purity results by HPLC-RP during 200 Cycles
Nessa Publishers| www.nessapublishers.com Page 15
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
Figure 10: Protein A leakage results during 200 cycles.
Figure 11: HCP results during 200 cycles.
During the lifetime test, no significant loss of purity is observed over the 200 cycles (figure 9). Protein A
leakage in the elution is stable as well, and even though a slight increase in HCP content is observed, it is
staying within the limits typically defined for this step for such attribute.
Furthermore, as a final observation and as it is mentioned in the experimental section of this article, the eluted
product is also analyzed between two NaOH CIP’s, for cycles 31, 33, 35, 37 and 39 whose CIP step is carried
out with the acidic PAB mixture only. No significant difference is observed regarding purity, HCP and protein
A leakage.
Nessa Publishers| www.nessapublishers.com Page 16
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
Conclusions
It was shown during the study described in this article that AbSolute HC Protein A media exhibits high
Dynamic Binding Capacities at low and high speed for Nimotuzumab capture. Beside this property of higher
loading, it was verified that product quality stays within specifications of purity, and is not impacted when the
speed of loading step, washing step or elution step is increased.
Both factors capacity and speed will allow to increase the productivity values to be reached with Protein A
media, expressed in terms of quantity of Nimotuzumab produced per Liter of resin per unit of time. Based on
those results, high performances in terms of productivity, purity, yield can be expected from an implementation
at larger scale of this capture step, with some stable results over 200 cycles under the conditions tested in this
study
Nessa Publishers| www.nessapublishers.com Page 17
Journal of Biotechnology and Bioengineering Volume 1| Issue 1
References
1. AbSolute® High Cap Revolutionary Protein A Media. Downstream Processing. BioProcess International
Industry. Yearbook 2012–2013.
2. Dynamic binding capacity study on MabSelectSuRe™ LX for capturing high-titer monoclonal antibodies.
GE Healthcare Life Sciences. Application note 28-9875-25 AA.
3. User Guide. AbSolute® High Cap by Novasep High Performance Protein A Media. Revision 0: November
2011.
4. Shukla A, Hubbard B, Tressel T, Guhan S, Low D. Downstream processing of monoclonal antibodies.
Application of platform approaches. J Chromatogr B. 2007; 848:28–39.
5. Swinnen K, Krul A, Van Goidsenhoven I, Van Tichelt N, Roosen A, Van Houdt K. Performance
comparison of protein A affinity resins for the purification of monoclonal antibodies. J Chromatogr B.
2007; 848:97–107.
6. G.Guiochon, A. Felinger, D.G. Shirazi, A.M. Katti, Fundamentals of Preparative and Nonlinear
Chromatography, 2nd ed., Academic Press, San Diego, 2006.
7. G. Guiochon, B. Lin, Modeling for Preparative Chromatography, Academic Press, New York, 2003.
8. Candy K.S. Ng, Hector Osuna-Sanchez, Eric Valéry, Eva Sorensen, Daniel G. Bracewell. Design of high
productivity antibody capture by Protein A chromatography using an integrated experimental and modeling
approach. J Chromatogr B. 2012; 899:116-126
9. R. Hahn, P. Bauerhansl, K. Shimahara, C. Wizniewski, A. Tscheliessnig, A. Jungbauer, J. Chromatogr. A
1093 (2005) 98.