Determination of residual poly diallyldimethylammonium chloride (pDADMAC) in monoclonal antibody formulations by size exclusion chromatography and evaporative light scattering detector

Biologicals xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Biologicals journal homepage: www.elsevier.com/locate/biologicals

Determination of residual poly diallyldimethylammonium chloride (pDADMAC) in monoclonal antibody formulations by size exclusion chromatography and evaporative light scattering detector Mehdi Khodadadian∗, Maryam Ghassemi, Hossein Behrouz, Shayan Maleknia, Fereidoun Mahboudi Biopharmaceutical Research Center, AryoGen Pharmed Inc., Alborz University of Medical Sciences, Karaj, Iran

A R T I C LE I N FO

A B S T R A C T

Keywords: Poly diallyldimethylammonium chloride Monoclonal antibody Size exclusion chromatography Evaporative light scattering detection

The cationic polyelectrolyte pDADMAC is widely used in biopharmaceutical industry as a flocculating agent to enhance clarification throughput and downstream filtration operations. Due to the possible toxicity, pDADMAC should be assessed for an acceptable residual level to ascertain the safety of the product to patients. The strong protein-polyelectrolyte interaction, however, can negatively affect sensitivity and accuracy of measurements. This paper reports on the application of size exclusion (SE) chromatography coupled to evaporative light scattering detector (ELSD) to the quantitative determination of pDADMAC in monoclonal antibody formulations and in process intermediates during downstream purification. The SE chromatography was performed under isocratic condition with a mobile phase consisting of 0.1% TFA in water (90%) and acetonitrile (10%) at a flow rate of 0.4 ml/min. A quantification limit (S/N = 10) of 0.85 ppm was achieved in sample matrix, which is sufficiently low for the trace analysis of this compound in protein-containing samples.

1. Introduction Biopharmaceuticals (or biologics) are drugs of biotechnological origin, which are used in the treatment of a variety of human diseases [1]. Most biologics are derived from non-human protein expression systems such as Chines Hamster Ovary (CHO) cells [2]. These include monoclonal antibodies (mAbs) and other recombinant protein products (e.g., fusion proteins, growth factors, cytokines, therapeutic enzymes, and hormones) as treatments for cancer, autoimmune diseases, and other life-threatening conditions [3]. Due to their high efficacy and specificity, the market for novel and biosimilar biopharmaceuticals is growing at a fast rate [4]. Because of advances in cell line engineering and cell culture media design, some studies have demonstrated cell culture processes that produce mAb titers as high as 25 g/L [5–7]. These high product titers are often associated with high cell densities and increased levels of contaminants from cell debris generated during cell culture and harvesting. This sub-micron cellular debris cannot be easily removed by existing solid-liquid separation techniques, as it results in the fouling of the downstream harvest filtration train. One way to overcome these challenges is flocculation of cell culture suspensions, which

∗

considerably enhances clarification throughput and downstream filtration operations [8,9]. Similar to other polycationic polymers used to flocculate cells, pDADMAC (Scheme 1) is known to be cytotoxic in vitro [10]. Thus, the removal of pDADMAC from process stream must be demonstrated by subsequent process operations to ensure clinical safety of the drug. The purification process must remove the residual polymer to a concentration less than 1 ppm where there is no in vitro cytotoxicity and hemolytic concerns [11–14]. Nevertheless, it might be difficult to accurately monitor pDADMAC levels at different stages during downstream processing since methods that detect and quantify this compound are mostly applied to potable water and not to complex solutions like cell culture broth or process intermediates [15–19]. Detection and quantification of pDADMAC in drug substances or final drug products with high protein content (as high as 185 mg/ml) is even more challenging because of the strong protein-polyelectrolyte interaction, which may severely affect sensitivity and accuracy of measurements [20]. On the other hand, pDADMAC bears no chromophore on its structure and is not electrochemically active. Furthermore, like other quaternary ammonium compounds, it underdogs few reactions and cannot be easily derivatized for UV or fluorescent detection.

Corresponding author. E-mail addresses: [email protected], [email protected] (M. Khodadadian).

https://doi.org/10.1016/j.biologicals.2018.11.002 Received 5 June 2018; Received in revised form 21 September 2018; Accepted 9 November 2018 1045-1056/ © 2018 International Alliance for Biological Standardization. Published by Elsevier Ltd. All rights reserved.

Please cite this article as: Khodadadian, M., Biologicals, https://doi.org/10.1016/j.biologicals.2018.11.002

Biologicals xxx (xxxx) xxx–xxx

M. Khodadadian et al.

2.4. Sample preparation 2.4.1. Samples with mAb concentration less than 60 mg/ml Caution: All the experiments must be performed under laboratory fume hood. Powder-free gloves and protective glasses must be used throughout the sample preparation. Four hundred microliters of protein sample is transferred into a 1.5 ml plastic microtube (Eppendorf). A same volume (400 μl) of 37% HCl (12.3 M) is added to the microtube to reach a final acid concentration of ∼6 M. Sample is then heated at 110 °C in a thermo-block for 6 h (Fig. 2S). Samples are removed from thermo-block and are allowed to cool down to room temperature. After centrifuging for a few seconds (on a micro-spin device), samples are placed into a centrifugal evaporator for 10–12 h. The dried sample is reconstituted with 50 μl of 0.5% TFA in water by vortexing for a few minutes to reach a homogeneous mixture (In the case of formation of a solid residue, a 10-μl micropipette tip may help redissolve the residue). The sample is then centrifuged at 12000 rpm for 5 min. Forty microliters of the supernatant is transferred into a 250-μl polypropylene vial (Agilent) for UPLC analysis (glass vials should not be used).

Scheme 1. Chemical structure of pDADMAC.

The evaporative light-scattering detector (ELSD) is the preferred detector for some applications such as carbohydrate, lipid and polymer analysis [21–23]. It is based on the amount of light scattered by particles in mobile phase that have been dried out through evaporation [24]. These detectors are considered universal and are used for the analysis of compounds where UV or fluorescent detection might be a restriction. The need for volatile mobile phases, low sensitivity, destructive detection and non-linear response are some limitations of this kind of detector [24]. In view of the challenges faced in the determination of pDADMAC in complex matrices especially in high protein content formulations, the study reported herein aimed at developing an analytical method for detection of residual pDADMAC in therapeutic monoclonal antibody formulations. To the best of our knowledge, this is the first report that exploits SE-UPLC-ELSD technique along with a simple sample preparation procedure for quantification of pDADMAC in biopharmaceutical formulations and in process intermediates.

2.4.2. Samples with mAb concentration greater than 60 mg/ml The same procedure is applied to protein formulations that are more concentrated. However, as the residue is so bulky that cannot be dissolved in a volume as small as 50 μl, a larger volume of the solvent should be used for reconstitution (e.g. 200 μl). The resultant homogenous mixture is subjected to ultrafiltration using Amicon® Ultra Centrifugal Filters (10-kDa cut off) to reduce the concentration of digestion products. This may be repeated two or three times to reach a less viscous sample whose volume can be easily reduced to ∼50 μl. The final solution is centrifuged at 12000 rpm for 5 min. Forty microliters of the supernatant is transferred into a 250-μl polypropylene vial for UPLC analysis.

2. Experimental section 2.1. Materials

2.5. Standard addition method

All mAb drug products and in-process samples were supplied by AryoGen Pharmed. Potassium hydroxide, 37% hydrochloric acid (HCl), HPLC grade acetonitrile (ACN), trifluoroacetic acid (TFA), 10% (w/w) pDADMAC solution all were purchased from Merck (Darmstadt, Germany). For solvent or buffer exchange, Amicon® Ultra Centrifugal Filters with 10-kDa cut-off were used.

Because negative samples (samples without pDADMAC) may not be always available to construct a calibration curve and ELSD response is not linear over a wide concentration range, pDADMAC concentration in an unknown sample can be estimated by standard addition method [25]. Given the concentration of added standard in final solution (Cs), area of sample (Ax) and area of sample after addition of standard (Ax+s), the unknown concentration (Cx) can be calculated using the following formula:

2.2. Instrumentation Chromatographic experiments in this study were performed on an Agilent 1290 Infinity II UPLC system equipped with a 1260 Infinity II ELSD detector. SE column was an AQUITY UPLC® Protein BEH SEC column (1.7 μm, 4.6 × 300 mm, Waters). Reversed phase (RP) experiments were performed on an Aeris WIDEPORE XB C18 column (3.6 μm, 4.6 × 150 mm, Phenomenex). Instrument control, data acquisition and compilation of results were performed using OpenLab software (Rev. C.01.07). A CHRIST centrifugal evaporator model RVC 2–25 CD-plus were used for drying digested samples.

Cx =

Cs

(

Ax +s Ax

)

−1

Here we assumed that the volume of sample (Vx) is not changed after addition of a small volume of standard (Vs). Example: Four hundred microliters of a protein sample is transferred into a 1.5 ml microtube (Vx = 400 μl, Cx = unknown). Two microliters of 1000-ppm pDADMAC standard solution is added to the same sample in another 1.5 ml microtube (Vs = 2 μl, Vx+s = 402 μl, Cs ≈ 5 ppm). Both samples are subjected to acid digestion and analyzed by SE-UPLCELSD technique as described in previous sections. The parameters Ax and Ax+s are simply determined by integrating the corresponding chromatograms of unknown and spiked samples. Finally, Cx is calculated using the above formula.

2.3. SE-UPLC-ELSD The mobile phase was a mixture of 0.1% TFA in water (90%) and acetonitrile (10%), which was isocratically pumped at a constant flow rate of 0.4 ml/min (column backpressure < 400 bar). The column was kept at room temperature and injection volume was set to 40 μl. The ELSD operated at a nebulizer temperature of 30 °C and an evaporation temperature of 65 °C. The nitrogen gas flowrate was 1.5 l/min (SLM). The PMT gain and smoothing factor were set to 1.0 and 30 (= 3 s), respectively. Before analysis, the column should be washed with mobile phase for at least 50 min to reach a steady baseline (Fig. 1S). The background noise should be not more than 0.2 mV.

3. Results and discussion 3.1. Reversed-phase chromatography of pDADMAC Reversed-phase HPLC on a C18 column with evaporative light 2

Biologicals xxx (xxxx) xxx–xxx

M. Khodadadian et al.

Fig. 1. RP chromatography of pDADMAC: a) Injection of a 5-ppm standard solution and b-e) four successive blank injections after the first injection.

Fig. 2. SE chromatography of pDADMAC: a) Injection of a 5-ppm standard solution and b) a blank injection after the first injection.

3.2. Size exclusion (SE) chromatography of pDADMAC

scattering detection is recommended by the manufacturer of pDADMAC (EMD Millipore) for the analysis of this compound in aqueous solutions [11]. However, the high carryover of pDADMAC on reversed-phase columns may significantly affect the accuracy and precision of the method. Fig. 1, as an example, shows that after injection of a 5-ppm pDADMAC standard solution onto the column at least four blank (water) injections are required to completely remove traces of pDADMAC. On the other hand, due to the strong hydrophobic interaction between the column and protein, protein-containing samples could not be injected directly. Our attempts, to develop a sample cleanup procedure before RP-HPLC-ELSD to eliminate matrix effect and to reach a low detection limit (< 1 ppm) all failed. Weak analytical response, analyte carryover and appearance of interfering peaks at/or around the retention time of pDADMAC were common problems associated with RP columns.

As an alternative, SE chromatography can be considered for the analysis of pDADMAC. Fig. 2a displays the SE chromatogram of a 5ppm pDADMAC standard solution. The pDADMAC signal emerges after five minutes. Solvent or any low molecular weight substance, which might be present in the sample, is eluted later at a retention time between 11 and 12 min. One important advantage of SE chromatography over reversed-phase chromatography is that analyte carryover is not observed in SE chromatography (Fig. 2b). Hence, extensive washing or numerous blank injections are not required to clean and regenerate the column. Analysis of a mAb drug product by SE chromatography, however, revealed that the retention time of protein aggregates (MW = 300 kDa) is very close to that of pDADMAC (MW = 350 kDa) (Fig. 3). The two overlapping peaks, in fact, could not be resolved on 3

Biologicals xxx (xxxx) xxx–xxx

M. Khodadadian et al.

Fig. 3. SE chromatograms of a) a mAb drug product (1 mg/ml) and b) 50-ppm pDADMAC standard solution.

Fig. 4. SE chromatograms of a) an acid digested mAb drug product (60 mg/ml) spiked with 5 ppm of pDADMAC and b) the same sample without pDADMAC. The injection volume was 2 μl in both cases.

4

Biologicals xxx (xxxx) xxx–xxx

M. Khodadadian et al.

Fig. 5. The use of a switching valve before ELSD to avoid detector contamination. SE chromatograms are belong to the same samples as in Fig. 4 with an injection volume of 40 μl.

excess acid can be easily evaporated and the dried sample can be reconstituted with an aqueous solution of TFA (0.5% v/v). To reach a lower detection limit, the sample can be reconstituted with a smaller volume of the solvent. As digested sample contains a high concentration of amino acids, we first injected a small volume (2 μl) of a digested protein sample (spiked with 5 ppm of pDADMAC) to have a general assessment of the resultant chromatogram (Fig. 4). The chromatogram shows that pDADMAC signal appears approximately two minutes earlier than a broad and intense peak, which is, in fact, due to the digestion products and formulation buffer ingredients. To avoid contamination of ELSD detector, the column outflow can be directed into the waste using a switching valve when a baseline reaches after pDADMAC signal (Fig. 5).

Table 1 Experimental variables (factors) and selected intervals in robustness testing. Variable

Central Value

Lower Value (−1)

Upper Value (+1)

A: Digestion temperature B: Incubation time C: TFA concentration for reconstitution D: TFA concentration in mobile phase E: Column temperature F: Flow rate G: Evaporation temperature (ELSD)

110 °C 6.0 h 0.50%

105 °C 5.7 h 0.47%

115 °C 6.3 h 0.53%

0.10%

0.095%

0.105%

25.0 °C 0.50 ml/min 65.0 °C

23.7 °C 0.47 ml/min 62 °C

26.3 °C 0.53 ml/min 68 °C

3.4. Validation of the method

this column and development of a sample cleanup procedure was indispensable to eliminate the interference of protein. As analyte carryover is not an issue in SE chromatography and low molecular weight substances do not interfere with determination of pDADMAC, SE chromatography was the method of choice for the analysis of pDADMAC in this study.

The proposed SE-UPLC-ELSD method for determination of pDADMAC was validated according to ICH Q2(R1) guideline [26] by determining its performance characteristics regarding repeatability, intermediate precision, linearity, specificity, limit of detection (LOD), limit of quantification (LOQ), accuracy, range and robustness. A negative drug product, with a mAb concentration of 60 mg/ml, was spiked with different concentrations of pDADMAC to evaluate validation parameters. Repeatability was checked by three separate sample preparations at three concentration levels (i.e. 0.85, 3.0 and 8.0 ppm) and subsequent SE-UPLC-ELSD analysis. RSDs less than 8.0% were obtained for peak area (Table 1S). Samples containing higher concentrations of pDADMAC or lower concentrations of protein, however, show more reproducible results. To evaluate intermediate precision the effects of three factors including day, analyst and column lot were studied using a 23−1 fractional factorial design matrix with four runs (Table 2S) (a general notation of 2k−p is used to denote fractional designs where k is the number of factors and p is the level of fractionation) [27]. An RSD of

3.3. Sample treatment Several protocols for the removal of proteins and their aggregates from sample were tried (e.g. enzymatic digestion, methanol precipitation, salting out, protein adsorption on cationic resins, etc.). The best and simplest approach was, indeed, acid/base digestion. Both acid and base digestion could be used for this purpose. The problem with base digestion (KOH 4 M, 110 °C), however, is that samples must be solvent exchanged several times (at least 7–8 times) before chromatographic analysis to remove excess KOH. To skip the exhausting solvent exchange step, we tried acid digestion with 6 M HCl at 110 °C for 6 h. The 5

Biologicals xxx (xxxx) xxx–xxx

M. Khodadadian et al.

Fig. 6. Half-normal probability plot for seven factors investigated in robustness study. A straight line is drawn through the non-significant effects.

Fig. 7. SE chromatograms of different samples analyzed by the proposed method. ELSD signals before (a) and after (b) addition of 5 ppm of pDADMAC standard to a drug product (60 mg/ml) are shown.

6

Biologicals xxx (xxxx) xxx–xxx

M. Khodadadian et al.

concentration of 60 mg/ml, whose initial cell broth was treated with 0.05% (w/v) pDADMAC (500 ppm) (Fig. 7). The concentration of pDADMAC was estimated as 0.37 ppm by standard addition method.

6.5% indicated that the variation is mainly due to random errors and selected factors does not significantly affect the response. Although ELSD does not necessarily produce a linear response over a wide concentration range, the results showed that there was a linear relationship between ELSD signal and pDADMAC concentration over a concentration range from 0.5 to 8.0 ppm with a correlation coefficient of 0.99 (Fig. 3S). The interference study revealed that the only interfering species are protein aggregates and, to a lower extent, protein itself, which can be completely removed from the solution by acid digestion. As ELSD shows a significant background noise, a signal-to-noise (S/N) approach was used to estimate LOD and LOQ of the method. The LOD (S/N = 3) and LOQ (S/N = 10) were estimated as 0.26 ppm and 0.85 ppm, respectively (Fig. 4S). To check the accuracy of the method negative samples were spiked with 0.85, 3.0 and 8.0 ppm of pDADMAC and the concentrations were determined (n = 3) by standard addition method. The mean recoveries ranged from 92 to 106% (Table 3S), which are acceptable for the intended purpose of the method. The range was determined from precision, accuracy and linearity studies. The analytical procedure showed an acceptable degree of precision (RSD < 8%) and accuracy (Recovery: 100 ± 10%) within the concentration range 0.85–8.0 ppm. According to ICH Q2(R1) “the robustness of an analytical procedure is a measure of its capacity to remain unaffected by small, but deliberate variations in method parameters and provides an indication of its reliability during normal usage”. Many ideas on robustness testing can be found in the literature, but the most useful approach is the application of experimental design [27–29]. Two-level screening designs like Plackett-Burman and fractional factorial designs have been the recommended designs for robustness testing [27,28]. Here, a 27−3 fractional factorial design was used to evaluate the robustness of the method. The selected factors, whose selection was based on our experience with this test, are listed in Table 1. The central values in the first column were altered by approximately ± 5% to define upper (+1) and lower (−1) levels of factors for the assessment of robustness. Table 4S is the plan of fractional factorial design to conduct the experiments. The design includes four center points, which are usually added to fractional designs with numeric factors to estimate pure error [30]. Experiments were performed according to Table 4S and the obtained responses were inserted in the last column of the table. An analysis of fractional factorial design by Half-normal plot (Fig. 6) clearly demonstrates that factor A (digestion temperature) and factor D (TFA concentration in mobile phase) are the two critical factors, which may significantly influence the response. Hence, these factors should be accurately set to their nominal values and precisely controlled during the experiments in order to mitigate or even eliminate any risk arising from the fluctuation of these factors. The method, however, is robust when other factors shift by not more than ± 5% from the midpoints.

3.6. Column lifetime The results showed that a brand-new column retains its initial performance after 300 injections of digested protein samples. It is estimated that the column can be used up to 400 injections. After prolonged usage a small carry over at the retention time of pDADMAC might be appeared. In this case the column can be reversed for another at least 100 injections. Periodic wash of SE column with 0.2 M KCl and then ultrapure water is recommended to increase the column lifetime. Moreover, injection of a monoclonal antibody solution (1.0 mg/ml) can effectively remove traces of pDADMAC from column. 3.7. Residual pDADMAC level in drug substance/products It has been demonstrated that a downstream mAb purification process that involves the two critical steps of Protein A chromatography and cation exchange chromatography can reduce pDADMAC level to concentrations less than 1 ppm (from a maximum initial concentration of 500 ppm (0.05% w/v)) [12–14]. Although this may be considered a successful clearance of the flocculant from in-process samples during downstream purification, the allowable level of pDADMAC in a mAb drug substance (or drug product) should be scientifically determined and justified. Manufacturers should refer to current applicable regulatory guidelines (e.g. ICH Q3A/ICH Q3B) to establish an acceptance criterion for this impurity in their products. The acceptance criterion should be justified with appropriate safety considerations. This work, however, provides a validated and suitable analytical procedure for detection and quantification of this impurity in drug substances or drug products. 3.8. Conclusion In this work, we described the quantitative determination of residual pDADMAC in therapeutic mAb formulations and in-process samples by SE chromatography and evaporative light scattering detection. Unlike reversed-phase chromatography, analyte carryover was not an issue in SE chromatography, which guaranteed good reproducibly of ELSD response. The acid digestion perfectly eliminated protein aggregates from solution, which severely interfered with determination of pDADMAC. The proposed method is sufficiently accurate, precise and sensitive for determination of residual pDADMAC in mAb drug substance/products and in-process samples. Acknowledgments

3.5. Application of the method This work was supported by Biopharmaceutical Research Center, AryoGen Pharmed Inc., Karaj, Iran and Alborz University of Medical Sciences (Ethics code: IR.ABZUMS.REC.1397.029).

The proposed analytical method was applied to analyze two inprocess samples during downstream purification and a finished drug product. After flocculation of the cell broth, flocculent settling and microfiltration of the clarified supernatant, the chromatographic purification includes separation of the target mAb on a Protein A column and the removal of cationic polymer on a sulphopropyl (SP) strong cation exchange resin. Eluates form Protein A and SP columns were analyzed by the proposed method for quantification of pDADMAC. The results are presented in Fig. 7. Surprisingly, a high level of pDADMAC was detected in Protein A eluate (∼86 ppm) indicating that Protein A chromatography imperfectly clears pDADMAC (initial pDADMAC concentration was 300 ppm). The analysis of SP eluate, however, reveals that the strong cation exchange resin removes the cationic polymer almost completely. The concentration of pDADMAC was reduced to levels far below the LOD of the method (< 0.26 ppm). The proposed method was also applied to the analysis of a drug product with a mAb

Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.biologicals.2018.11.002. References [1] Jungbauer A, Graumann K. Biopharmaceuticals – discovery, development and manufacturing. Biotechnol J 2012;7:1422–3. [2] Kim JY, Kim YG, Lee GM. CHO cells in biotechnology for production of recombinant proteins: current state and further potential. Appl Microbiol Biotechnol 2012;93:917–30. [3] Ho RJY, Gibaldi M. Biotechnology and biopharmaceuticals: transforming proteins and genes into drugs. second ed. John Wiley & Sons; 2013.

7

Biologicals xxx (xxxx) xxx–xxx

M. Khodadadian et al.

2006;1101:222–5. [17] Mwangi IW, Ngila JC, Ndungu P, Msagati TAM. Method development for the determination of diallyldimethylammonium chloride at trace levels by epoxidation process. Water Air Soil Pollut 2013;224:1638–46. [18] Mwangi IW, Ngila JC, Ndungu P. A new spectrophotometric method for determination of residual polydiallyldimethylammonium chloride flocculant in treated water based on a diazotization-coupled ion pair. WaterSA 2012;38:707–14. [19] Ziółkowska D, Shyichuk A, Żelazko K. A new spectrophotometric method for determination of residual polydiallyldimethylammonium chloride flocculant in treated water based on a diazotization-coupled ion pair. Polimery 2012;4:303–5. [20] Kayitmazer AB, Seeman D, Minsky BB, Dubin PL, Xu Y. Protein–polyelectrolyte interactions. Soft Matter 2013;9:2553–83. [21] Condezo-Hoyos L, Pérez-López E, Rupérez P. Improved evaporative light scattering detection for carbohydrate analysis. Food Chem 2015;180:265–71. [22] Perona JS, Ruiz‐Gutierrez V. Quantification of major lipid classes in human triacylglycerol‐rich lipoproteins by high‐performance liquid chromatography with evaporative light‐scattering detection. J Separ Sci 2004;27:653–9. [23] Arndt JH, Macko T, Brüll R. Application of the evaporative light scattering detector to analytical problems in polymer science. J Chromatogr A 2013;1310:1–14. [24] Megoulas NC, Koupparis MA. Twenty years of evaporative light scattering detection. Crit Rev Anal Chem 2005;35:301–16. [25] Bader M. A systematic approach to standard addition methods in instrumental analysis. J Chem Educ 1980;57:703–6. [26] ICH Q2 (R1). Validation of analytical procedures: text and methodology. 2005. [27] Ragonese R, Mulholland M, Kalman J. Full and fractionated experimental designs for robustness testing in the high-performance liquid chromatographic analysis of codeine phosphate, pseudoephedrine hydrochloride and chlorpheniramine maleate in a pharmaceutical preparation. J Chromatogr A 2000;870:45–51. [28] Dejaegher B, Dumarey M, Capron X, Bloomfield MS, Vander Heyden Y. Comparison of Plackett–Burman and supersaturated designs in robustness testing. Anal Chim Acta 2007;595:59–71. [29] Vander Heyden Y, Nijhuis A, Smeyers-Verbeke J, Vandeginste BG, Massart DL. Guidance for robustness/ruggedness tests in method validation. J Pharmaceut Biomed Anal 2001;24:723–53. [30] Montgomery DC. Design and analysis of experiments. eighth ed. John Wiley & Sons; 2012.

[4] Blackstone EA, Fuhr Jr. JP. Biopharmaceuticals: the economic equation. Biotechnol Healthc 2007;4:41–5. [5] Schirmer EB, Kuczewski M, Golden K, Lain B, Bragg C, et al. Primary clarification of very high-density cell culture harvests by enhanced cell settling. BioProcess Int 2010;8:32–9. [6] Chon JH, Papastoitsis GZ. Advances in the production and downstream processing of antibodies. N Biotech 2011;28:458–63. [7] Butler M, Acosta AM. Recent advances in technology supporting biopharmaceutical production from mammalian cells. Appl Microbiol Biotechnol 2012;96:885–94. [8] Singh N, Arunkumar A, Chollangi S, Tan ZG, Borys M, Li ZJ. Clarification technologies for monoclonal antibody manufacturing processes: current state and future perspectives. Biotechnol Bioeng 2016;113:698–716. [9] McNerney T, Thomas A, Senczuk A, Petty K, Zhao X, Piper R, et al. PDADMAC flocculation of Chinese hamster ovary cells: enabling a centrifuge-less harvest process for monoclonal antibodies. mAbs 2015;7:413–27. [10] Fischer D, Li Y, Ahlemeyer B, Krieglstein J, Kissel T. In vitro cytotoxicity testing of polycations: influence of polymer structure on cell viability and hemolysis. Biomaterials 2003;24:1121–31. [11] Application Note. pDADMAC flocculant reagent for use with Clarisolve® depth filters. Billerica, MA USA: EMD Millipore Corporation; 2015. [12] Burgstaller D, Krepper W, Haas J, Maszelin M, Mohoric J, Pajnic K, et al. Continuous cell flocculation for recombinant antibody harvesting. J Chem Technol Biotechnol 2018;93:1881–90. [13] Felo M, Kang Y, Hamzik J, Balderes P, Ludwig DL. Industrial application of impurity flocculation to streamline antibody purification processes. Pharm Bioprocess 2015;3:115–25. [14] Kang Y, Hamzik J, Felo M, Qi B, Lee J, Ng S, et al. Development of a novel and efficient cell culture flocculation process using a stimulus responsive polymer to streamline antibody purification. Biotechnol Bioeng 2013;110:2928–37. [15] Trisaranakul W, Chompoosor A, Maneeprakorn W, Nacapricha D, Choengchan N, Teerasong S. A simple and rapid method based on anti-aggregation of silver nanoparticles for detection of poly (diallyldimethylammonium chloride) in tap water. Anal Sci 2016;32:769–73. [16] Jin F, Hu J, Yang M, Jin X, He W, Han H. Determination of diallyldimethylammonium chloride in drinking water by reversed-phase ion-pair chromatography–electrospray ionization mass spectrometry. J Chromatogr A

8