Protein Expression and Purification 164 (2019) 105460

Contents lists available at ScienceDirect

Protein Expression and Purification

journal homepage: www.elsevier.com/locate/yprep

Sodium caprylate induced precipitation post Protein A chromatography as T

an effective means for host cell protein clearance
Yan Wan, Ting Zhang, Tao Chen, Ying Wang, Yifeng Li*
Technology and Process Development (TPD), WuXi Biologics, 288 Fute Zhong Road, Waigaoqiao Free Trade Zone, Shanghai, 200131, China

A R T I C LE I N FO A B S T R A C T

Keywords: In downstream processing of monoclonal antibody (mAb), post Protein A neutralization and subsequent inter-
CA mediate depth filtration are critical steps for host cell protein (HCP) clearance. Previous studies have shown that
Caprylic acid adding caprylic acid (CA) during neutralization can further improve HCP removal by promoting their pre-
HCP cipitation. In this study, we replaced CA with its sodium salt – sodium caprylate (SC). For the five mAbs studied,
Host cell protein
SC has been shown to be equally effective as CA at precipitating HCPs. As the salt form has a higher solubility, SC
mAb
Monoclonal antibody
stock solution with relatively high concentration can be easily prepared, which facilitates its adding to the
Protein A eluate Protein A elution pool. Thus, this study not only confirms the effectiveness of CA/SC-induced HCP precipitation
SC but also provides a more convenient way to integrate this method into the downstream process.
Sodium caprylate

1. Introduction which in turn exposes additional binding sites. When CA binding

In a typical platform for monoclonal antibody (mAb) purification,
Protein A chromatography is followed by low pH viral inactivation and
then the product pool is neutralized to a pH suitable for the next
chromatography step. Previous studies have shown that pH adjustment
of Protein A eluate or the viral inactivated pool is often accompanied
with precipitation and maximum turbidity usually occurs between pH
5.0 and 7.5 [1]. Moreover, precipitation was found to be strongly
correlated with selective removal of host cell proteins (HCPs) [2]. The
rationale behind this phenomenon is that most Chinese hamster ovary
(CHO) cell HCPs have isoelectric points (pIs) between 4.5 and 7.0, and
become less soluble when titrated to this pH range [2,3]. The induced
precipitates are usually removed by depth filtration [4,5]. In general,
neutralization post Protein A and the subsequent intermediate depth
filtration are critical steps for HCP clearance in downstream processing
of mAbs.
Caprylic acid (CA) is a naturally occurring eight-carbon fatty acid,
which is known to be able to selectively precipitate non-IgG proteins
[6,7]. CA, which contains the hydrophobic alkyl chain and the hydro-
philic carboxylate, is an amphipathic molecule and can function as a
surfactant [8–10]. It has been proposed that CA binds to specific sites of
the precipitating protein, inducing partial unfolding of the protein,
reaches certain threshold, it triggers protein denaturation and pre-
cipitation [11]. For protein solution at a given concentration, a
minimum amount of CA is required to start precipitation and this
triggering CA concentration varies for different proteins [11]. In addi-
tion, for the same protein this thresholding CA amount increases with
protein concentration. These characteristics form the basis of CA
mediated selective precipitation of proteins with different properties
and/or concentrations (e.g., hydrophobic and low-concentration pro-
teins will be precipitated first at a fixed CA amount). Brodsky et al.
recently showed that adding CA at the neutralization stage post Protein
A chromatography can further promote selective precipitation of HCPs
[12]. The authors studied CA-induced precipitation post Protein A with
six mAbs. It was found that the optimal pHs for selective HCP pre-
cipitation fell within a range of 5.0–5.4 at a fixed CA concentration of
1%. HCP removal ranged from 93% to 100%. In another work, Zheng
et al. conducted similar studies with five mAbs [13]. In consistent with
Brodsky et al.’s results, the authors also leaned while the optimal con-
dition varied slightly for different antibodies, an operational pH range
(5.0–6.0) and CA concentration (0.5–1.0%) were identifid for all five
mAbs. For four out of the five mAbs, HCP reduction was greater than
99% and for the other mAb reduction reached 79%.
CA is an oily liquid that is minimally soluble in water and has a

Abbreviations: mAb, monoclonal antibody; HCP, host cell protein; CHO, Chinese hamster ovary; pI, isoelectric point; CA, caprylic acid; SC, sodium caprylate; SEC-
HPLC, size-exclusion chromatography-high performance liquid chromatography
*
Corresponding author. 288 Fute Zhong Road, Waigaoqiao Free Trade Zone, Shanghai, 200131, China.
E-mail address: li_yifeng@wuxiapptec.com (Y. Li).

https://doi.org/10.1016/j.pep.2019.105460
Received 26 June 2019; Received in revised form 23 July 2019; Accepted 23 July 2019
Available online 25 July 2019
1046-5928/ © 2019 Elsevier Inc. All rights reserved.
Y. Wan, et al.
slightly unpleasant rancid-like smell. In the above-mentioned two stu-
dies 2% CA emulsion or neat CA was added to the Protein A elution
pool to reach a final CA concentration of 0.5–1.0% (v/v) [12,13]. In
either case, preparing 2% CA emulsion or transferring oily CA liquid is
not convenient. The sodium salt of CA, sodium caprylate (SC), displays
much higher solubility than CA. In this study, we replaced CA with SC
in treating Protein A elution pool and confirmed that under appropriate
conditions SC is equally effective at selective precipitating HCPs. In
comparison with CA, SC is much easier to handle. Thus, replacement of
CA with SC provides a more convenient way to integrate this method
into the downstream process.
2. Materials and methods
2.1. Materials
SC, sodium acetate trihydrate, sodium chloride, sodium hydroxide,
Tris base and ethanol were purchased from Merck (Darmstadt,
Germany). Acetic acid were purchased from J.T. Baker (Phillipsburg,
NJ, USA). D0HC and A1HC were purchased from Merck (Darmstadt,
Germany). MabSelect SuRe LX and the HiScale 50 column (5.0 cm I.D.)
were obtained from GE Healthcare (Uppsala, Sweden). The five mAbs
used in this study were expressed in CHO–K1 cells. The titers at harvest
for different antibodies range from 2 to 5 g/L.
2.2. Equipment
Protein A chromatography runs were conducted on an AKTA pure
150 system installed with Unicorn software version 7.3 (GE Healthcare,
Uppsala, Sweden). pH and conductivity was measured using
SevenExcellence S470 pH/Conductivity meter (Mettler-Toledo,
Columbus, OH, USA). Clarification of small amount samples was per-
formed using Centrifuge 5418 (Eppendorf, Mittelsachsen, Saxony,
Germany). Protein concentration was measured using a NanoDrop one
spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). An
Agilent 1260 liquid chromatography instrument (Agilent Technologies,
Santa Clara, CA, USA) was used for size-exclusion chromatography-high
performance liquid chromatography (SEC-HPLC) analysis. The plate for
HCP quantitation was read using Infinite 200 PRO plate reader from
Tecan (Männedorf, Switzerland).
2.3. Methods
2.3.1. Protein a chromatography
MabSelect SuRe LX was packed in a 5.0 cm diameter column with
23 cm bed height. The column was loaded at densities ranging from 30
to 50 mg protein per ml of resin for different antibodies. After loading,
the column was washed consecutively with 50 mM Tris-Acetate,
150 mM NaCl, pH 7.4, 50 mM Na-Acetate, 1 M NaCl, pH 5.5 and 50 mM
Na-Acetate/HAc, pH 5.5, each for 5 CV. The column was eluted with
50 mM Na-Acetate/HAc, pH 3.5.
2.3.2. SC induced precipitation
Protein A eluates of different antibodies were adjusted to 20 mg/ml
through dilution or concentration (when the pool concentration is
above or below 20 mg/ml, respectively), and in each case this con-
centration adjusted Protein A elution pool was used as the starting
material for SC precipitation. A 0.5 M SC stock solution was prepared by
dissolving SC in water. For condition screening, SC stock solution was
added to small aliquot (3–5 ml) of samples to reach different final
concentrations. After adding SC, the samples were pH adjusted to 5.0 or
5.5 with 1 M Tris base. After SC adding and pH adjustment, the samples
became turbid. Precipitates were removed by high speed centrifugation
(10,000×g, 2 min), and clarified samples were subjected to con-
centration, HCP measurements and SEC-HPLC analysis. In each case, a
suitable condition was selected for scale up based on recovery (HCP and
2
Protein Expression and Purification 164 (2019) 105460
SEC data were not available yet when making the selection). In general,
reduced yield/increased turbidity suggest that more HCPs were pre-
cipitated (when HCPs precipitate, a portion of the product was in-
evitably coprecipitated, resulting in dropped yield). In this study, the
highest SC concentration that gave a close to 90% yield was selected for
scale-up for each mAb. For the scale-up experiments, in each case
200–500 ml of Protein A elution pool (pH 4.0–4.4) was treated with the
pre-determined amount of SC and then adjusted to the selected pH.
Precipitates were removed by filtering the samples through an A1HC
depth filter loaded at densities ranging from 1500 to 2000 g/m2.
2.3.3. SEC-HPLC
SEC-HPLC analysis was performed on an Agilent 1260 liquid chro-
matography instrument using a Tosoh TSKgel G3000SWxl stainless
steel column (7.8 × 300 mm). 100 μg of sample was injected per run.
The mobile phase consisted of 50 mM sodium phosphate, 300 mM so-
dium chloride at pH 6.8. Each sample was eluted isocratically for
20 min at a flow rate of 1.0 ml/min. Protein elution was monitored by
UV absorbance at 280 nm. The peaks corresponding to aggregates,
monomer and low-molecular-weight species were integrated to calcu-
late the percentage of each species.
2.3.4. HCP measurement
HCP level in Protein A eluate (with or without SC treatment) was
measured using the 3rd generation generic ELISA kit from Cygnus
Technologies (Southport, NC, USA) following manufacturer's instruc-
tions. The detection range is 3–100 ng/ml. Serial dilutions of samples
were made to keep the measurement within the calibration range.
Absorbance was measured at 450 nm (absorbance) and 630 nm (re-
ference) using Infinite 200 PRO plate reader (Tecan, Männedorf,
Switzerland).
3. Results and discussion
3.1. Protein a chromatography
Protein A chromatography was performed following standard pro-
tocol as described in the Methods section. A representative chromato-
gram from mAb A is shown in Fig. 1. Protein A elution pool pHs and
HCP levels of different antibodies range from 4.0 to 4.4 and 1000 to
6000 ppm, respectively.
3.2. SC induced precipitation
For mAbs A-C and E, a portion of concentration adjusted Protein A
elution pool (pH 4.4, 4.3, 4.1 and 4.4, respectively) was aliquoted into
small-volume aliquots (3–5 ml). To find the most suitable SC con-
centration for HCP precipitation, the samples were treated with dif-
ferent amounts of SC (from 15/20 mM to 40/45 mM in increments of
5 mM) followed by pH adjustment to 5.5. For mAb D, concentration
adjusted Protein A elution pool (pH 4.0) was similarly aliquoted. In this
case, only two SC concentrations (i.e., 7.5 mM and 10 mM) were tested
due to observation of significant recovery drop at 10 mM SC. Also, in
this case pH was adjusted to 5.0 after adding SC. The relevant data of SC
concentration screening studies including yield, HCP and SEC purity
under different conditions for the five mAbs are summarized in Table 1.
For all antibodies except for mAb C, adjusting pH alone significantly
reduced HCPs. In the case of mAb C, adjusting pH alone only slightly
reduced HCP level. Nevertheless, for all five mAbs studied, SC treat-
ment induced further impurity precipitation, which greatly improved
HCP clearance. Appreciable difference was found among samples
treated with different amounts of SC and turbidity increased with in-
creasing amount of SC (Fig. 2). Furthermore, increase in turbidity
correlates well with HCP reduction (Table 1). Whereas HCP removal is
more effective at higher SC concentrations, product recovery dropped
rapidly at increased SC amounts. In all cases except for mAb C, SEC
Y. Wan, et al. Protein Expression and Purification 164 (2019) 105460

Fig. 1. A representative Protein A chromatogram from mAb A. The X-axis, left Y-axis and right Y-axis stand for volume (L), UV absorbance at 280 nm (mAU) and
conductivity (mS/cm), respectively.

Table 1 purity was not affected by SC treatment. For mAb C, adding SC caused
Recovery and quality data of SC condition screening experiments for five mAbs. significant increase in aggregation. Thus, in each case an optimal
mAb No. SC (mM) pH Recovery (%) HCP (ppm) SEC (%)
amount of SC should be selected by balancing HCP removal and pro-
duct recovery/SEC purity. In this study, as SEC analysis and HCP
A 1 0 4.4 100.0 4082 97.2 measurement takes longer time and the data were not immediately
2 0 5.5 100.0 845 96.9 available, SC concentration selected for scale-up was solely based on
3a 15 5.5 93.5 135 97.2
recovery data.
4 20 5.5 87.1 92 97.7
5 25 5.5 83.2 86 98.1 For mAbs A-D, SC treatment was performed under the selected
6 30 5.5 74.8 NAb NA condition at increased scale (200–500 ml). For mAb E, the scale-up
7 35 5.5 68.3 NA NA experiment was not performed due to insufficient amount of material.
8 40 5.5 57.5 NA NA
The relevant data of scale-up experiments are shown in Table 2. In
B 1 0 4.3 100.0 1162 94.1
2 0 5.5 99.9 583 93.9 general, data obtained at small and large scales (3–5 ml vs. 200–500 ml)
3 20 5.5 99.3 NA NA are highly consistent (Tables 1 and 2). The HCP levels of Protein A
4 25 5.5 96.9 NA NA elution pools for mAbs A-D are 4082 ppm, 1162 ppm, 6311 ppm and
5 30 5.5 95.7 80 94.0 2535 ppm, respectively. After SC treatment and depth filtration, the
6 35 5.5 94.0 54 94.3
corresponding values were reduced to 75 ppm, 29 ppm, 510 ppm and
7 40 5.5 91.4 11 95.1
8 45 5.5 87.3 10 95.6 37 ppm. Thus, 98.2%, 97.5%, 91.9% and 98.5% HCPs were removed
C 1 0 4.1 100.0 6311 96.6 respectively in these four cases. In normal procedures without SC
2 0 5.5 100.0 5505 96.7 treatment, depth filtration contributes greatly to HCP clearance as this
3 15 5.5 100.0 NA NA
step removes not only precipitated but also soluble HCPs. Under the
4 20 5.5 95.7 1621 94.0
5 25 5.5 96.7 NA NA
current condition, depth filtration made limited further contribution to
6 30 5.5 97.9 941 82.3 HCP clearance, suggesting that most HCPs were precipitated upon SC
7 35 5.5 89.7 446 77.8 treatment. For mAb C, despite effective HCP clearance, low product
8 40 5.5 83.4 308 79.7 recovery and SEC purity were observed under the selected condition,
D 1 0 4.0 100.0 2535 97.2
suggesting that a more conservative condition (e.g., 20 mM SC) should
2 0 5.0 100.0 903 96.7
3 7.5 5.0 96.3 129 98.3 be used. For mAbs A, B and D, the accumulative yields of SC treatment
4 10 5.0 86.8 47 98.6 and depth filtration are 89.1%, 93.1% and 81.9%, respectively.
E 1 0 4.4 100.0 2210 91.0 Whereas the yields are slightly lower than what is normally seen in
2 0 5.5 99.7 676 91.3
process without SC treatment, the new procedure potentially enables a
3 15 5.5 101.6 NA NA
4 20 5.5 101.5 NA NA
two-column downstream process considering SC's strong HCP-pre-
5 25 5.5 100.3 828 90.7 cipitating and virus-inactivating effects [12–14]. If the SC treatment
6 30 5.5 100.9 NA NA can indeed save one column chromatography, then the sacrifice in yield
7 35 5.5 100.0 576 91.4 at this step is worth it.
8 40 5.5 91.0 9 92.7
Besides SC concentration, pH is another parameter for effective HCP
9 45 5.5 86.0 2 93.1
precipitation by SC treatment. As mentioned in the previous sections,
a
For mAbs A-D, the condition selected for scale up are shown in bold and adjusting pH alone (in the range of 5.0–7.5) triggers HCP precipitation
italic. as the pIs of most CHO HCPs fall into that range. The pKa of CA is 4,89
b
Not available. and it is the nonionized form of CA that binds and precipitates proteins
[15]. At pH above pKa, the ionized form dominants and therefore the
amount of effective component is low. At pH below pKa, the nonionized
form dominants but the amount in solution may not increase due to this

3
Y. Wan, et al. Protein Expression and Purification 164 (2019) 105460

Fig. 2. Photo images of Protein A elution pool treated with different amounts of SC during the condition screening study. A-E, mAb A-E, respectively. Left 1,
Protein A eluate; left 2, pH adjusted Protein A eluate; the rest, pH adjusted Protein A eluate with various amounts of SC as indicated in Table 1. The SC concentrations
in mM are labelled on bottle cap. In all cases, appreciable difference was noticed among samples with different amounts of SC and turbidity increased with increasing
amount of SC.

Table 2 plasma by CA was best obtained at pH 5.0. In addition, when Brodsky

Recovery and quality data of scale-up experiments for four mAbs. et al. treated the Protein A elution pools of six mAbs with 1% CA, they
mAb SC (mM) pH Stage Recovery (%) HCP (ppm) SEC (%) found that the optimal pHs for HCP precipitation fell within a range of
5.0–5.4 [12]. As previous studies showed that pH has a less impact than
A 15 5.5 Aa 93.8 182 NAb SC concentration and the optimal pH falls within a narrow range, we
Bc 95.0 75d 97.7
selected pH 5.5 for convenience. However, for mAb D, significant yield
B 35 5.5 A 96.1 52 94.1
B 96.9 29 94.6
drop was observed at this pH and therefore pH 5.0 was used instead.
C 35 5.5 A 91.8 545 74.4 For different mAbs, fine tuning of pH is likely required for best results.
B 85.0 510 84.6
D 10 5.0 A 82.1 171 98.3
B 99.8 37 98.4 4. Conclusion
a
Post SC treatment.
b
In this study, we added SC to the Protein A elution pools of five
Not available.
mAbs to promote HCP precipitation. Under appropriate conditions
c
Post depth filtration.
d
The corresponding HCP levels in Protein A elution pool for mAb A-D are
(suitable pH and SC concentration), SC treatment significantly im-
4082 ppm, 1162 ppm, 6311 ppm and 2535 ppm, respectively. proved HCP removal while maintained reasonable product yield. After
the suitable condition was identified for each mAb at small scale, the
form's low water solubility. Thus, the best pH for CA/SC-induced pre- effectiveness of this approach was confirmed at increased scale.
cipitation should be around CA's pKa. In consistent with this conclusion, Previous studies used CA for the same purpose. However, CA's low
a previous study showed that precipitation of non-IgG fraction from solubility makes its use less convenient. SC, on the other hand, has a
much higher solubility and therefore is easier to handle. Thus, the

4
Y. Wan, et al.
current study not only confirms the effectiveness of CA/SC-induced
HCP precipitation but also provides a more convenient way to integrate
this method into the downstream process.
Acknowledgements
We would like to thank Weichang Zhou and Peter (Keqiang) Shen
for lending their support to this work. We thank Technology and
Process Development (TPD) Department's cell culture team and analy-
tical team for material generation and HCP measurement, respectively.
References
[1] S. Chollangi, R. Parker, N. Singh, Y. Li, M. Borys, Z. Li, Development of robust
antibody purification by optimizing protein-A chromatography in combination with
precipitation methodologies, Biotechnol. Bioeng. 112 (2015) 2292–2304.
[2] M. Jin, N. Szapiel, J. Zhang, J. Hickey, S. Ghose, Profiling of host cell proteins by
two-dimensional difference gel electrophoresis (2D-DIGE): implications for down-
stream process development, Biotechnol. Bioeng. 105 (2010) 306–316.
[3] J. Pezzini, G. Joucla, R. Gantier, M. Toueille, A.M. Lomenech, C. Le Sénéchal,
B. Garbay, X. Santarelli, C. Cabanne, Antibody capture by mixed-mode chromato-
graphy: a comprehensive study from determination of optimal purification condi-
tions to identification of contaminating host cell proteins, J. Chromatogr. A 1218
(2011) 8197–8208.
[4] Y. Yigzaw, R. Piper, M. Tran, A.A. Shukla, Exploitation of the adsorptive properties
of depth filters for host cell protein removal during monoclonal antibody
Protein Expression and Purification 164 (2019) 105460
purification, Biotechnol. Prog. 22 (2006) 288–296.
[5] J. Schreffler, M. Bailley, T. Klimek, P. Agneta, W.E. Wiltsie, M. Felo, P. Maisey,
X. Zuo, E. Routhier, Characterization of postcapture impurity removal across an
adsorptive depth filter, Bioproc. Int. 13 (2015) 36–45.
[6] M. Steinbuch, R. Audran, The isolation of IgG from mammalian sera with the aid of
caprylic acid, Arch. Biochem. Biophys. 134 (1969) 279–284.
[7] C. Russo, L. Callegaro, E. Lanza, S. Ferrone, Re.: purification of IgG monoclonal
antibody by caprylic acid precipitation, J. Immunol. Methods 65 (1983) 269–271.
[8] J.B. Rosenholm, The structure and properties of medium-chain surfactant solutions:
a case study of sodium octanoate, Adv. Colloid Interface Sci. 4l (1992) 197–239.
[9] M. Salavati-Niasari, F. Davar, T. Mahmoudi, A simple route to synthesize nano-
crystalline nickel ferrite (NiFe2O4) in the presence of octanoic acid as a surfactant,
Polyhedron 28 (2009) 1455–1458.
[10] L. Zhenga, B. Cuia, N.G. Akdogana, W. Lia, G.C. Hadjipanayis, Influence of octanoic
acid on SmCo5 nanoflakes prepared by surfactant-assisted high-energy ball milling,
J. Alloy. Comp. 504 (2010) 391–394.
[11] V. Morais, H. Massaldi, A model mechanism for protein precipitation by caprylic
acid: application to plasma purification, Biotechnol. Appl. Biochem. 59 (2012)
50–54.
[12] Y. Brodsky, C. Zhang, Y. Yigzaw, G. Vedantham, Caprylic acid precipitation method
for impurity reduction: an alternative to conventional chromatography for mono-
clonal antibody purification, Biotechnol. Bioeng. 109 (2012) 2589–2598.
[13] J. Zheng, L. Wang, B. Twarowska, S. Laino, C. Sparks, T. Smith, R. Russell, M. Wang,
Caprylic acid-induced impurity precipitation from protein A capture column elution
pool to enable a two-chromatography-step process for monoclonal antibody pur-
ification, Biotechnol. Prog. 31 (2015) 1515–1525.
[14] Y. Li, The application of caprylic acid in downstream processing of monoclonal
antibodies, Protein Expr. Purif. 153 (2019) 92–96.
[15] H.B. Bull, K. Breese, Binding of fatty acids by proteins, Arch. Biochem. Biophys. 120
(1967) 303–308.