Tangential Flow Filtration for the Recovery of Acellular Pertussis Vaccine Components

October 1, 2009

Membrane-based TFF technology can ease scale-up and provide a higher recovery percentage compared to conventional purification methods.


The commercial preparation of acellular pertussis vaccine depends on the effective recovery and purification of the antigens pertussis toxin (PT), filamentous haemaglutinin (FHA), and pertactin (PRN) from Bordetella pertussis fermentation. This study describes the recovery of the antigens using an open channel 0.45-μm tangential flow filtration module with optimization of the process parameters of transmembrane pressure, cross flow, and flux. Under the optimized conditions, greater than 98% recovery of FHA and PT was obtained.

Pertussis or whooping cough is an acute infectious disease of the respiratory tract caused primarily by Bordetella pertussis and less commonly by Bordetella parapertussis.1 Pertussis still continues to cause significant morbidity and mortality globally. "Per" meaning intensive and "tussis" cough: describes the clinical manifestation of the disease characterized by coughing and minor systemic complaints predominant in children.2B. pertussis is a nonmotile gram-negative coccobaccilli less than 1 μm in width and length. Diptheria toxoid and tetanus toxoid combined with whole-cell pertussis (DTP vaccine) has long been in use for childhood immunization programs. Whole-cell DTP vaccines are associated with local adverse reactions like erythrema, swelling, fever, and mild systemic events, e.g., convulsions and hypotonic hyporesponsive events.3 The reactogenecity of the whole cell pertussis vaccine has evoked public controversy in several countries and prompted the development of the acellular pertussis vaccine, which has a lower rate of adverse events and is effective in preventing pertussis.


Pathogensis of pertussis is a complex process and not completely understood. It includes pertussis toxin, filamentous hemagglutinin, pertactin, tracheal colonization factor (TCF), serum bactericidal resistance factor (BrkA), tracheal cytotoxin (TCT), adenyl cyclase toxin-hemolysin (ACT), lipoplooligosaccharide (LOS), heat labile toxin (HLT), and fimbriae 2 and 3 (Fims 2 & 3).4 Acellular pertussis vaccines contain inactivated pertussis toxin (PT), 105 kD, and may contain one or more of the bacterial components: filamentous hemagglutinin (FHA), 220 kD, a 69 kD outer membrane protein pertactin (PRN), and Fims 2 & 3.

To keep the disease under control large amounts of PT, FHA, and PRN must be prepared for mass immunization. The first step toward purification is the clarification of the fermentation broth for the recovery and further downstream processing of the antigens. The traditional method available is centrifugation. The scale-up of the process using centrifugation may lead to lower recovery of the antigens and a huge capital investment with low scalability. Therefore, an efficient bioseparation technique that can be easily scaled up with lesser capital investment is necessary. Membrane-based technology for biopharmaceuticals has been widely suggested and used.5,6 Tangential flow filtration (TFF) technology using an open channel, steam-sterilizable device allows microorganisms to be washed and concentrated in an efficient, rapid manner by using a closed recirculating filtration system, where both the concentrated cell mass and the filtrate can be collected under aseptic conditions. The factors that must be considered for successful scale-up of TFF technology are purity, yield, and reproducibility, for which the process parameters of the TFF must be optimized accurately. This article describes the optimization of the TFF process for high recovery of the acellular pertussis vaccine antigenic components: PT and FHA.

Materials and Methods

Bacterial Strains

A wild type B. pertussis strain Tohama-I was used in this study. The strain was procured from the American Type Culture Collection (Manassas, VA) as a freeze-dried ampoule.


Purified pertussis toxin, purified FHA, anti-PT serum (mouse), anti-FHA serum (mouse), anti-PT serum (sheep), and anti-FHA serum (sheep) were obtained from National Institute for Biologicals and Control (Hertfordshire, UK). Anti-mouse polyvalent immunoglobins peroxidase conjugate (GAMHRP) was obtained from Sigma (St. Louis, MO). Sodium-L-glutamate, glutamic acid, and CaCl2.2H2O were from Merck. All other chemicals were obtained from Sigma.

Growth Conditions

The B. pertussis Tohama strain was grown and maintained on Bordet-Gengou (B.G.) agar medium containing defibrinated horse blood. The strain was stored in a lyophilized state at 2–8 °C until grown on B.G. agar medium. The solid bacterial growth on B.G. agar slant, incubated for 48 h at 35 °C, was transferred into 500-mL Erlynmeyer flasks containing 200 mL of liquid medium similar to the defined medium described by Stainer and Scholte.7 One liter of medium contained 10.7 g of monosodium glutamate, 0.24 g of L-proline, 2.5 g of NaCl, 0.5 g of KH2PO4, 0.2 g of KCl, 0.1 g of MgCl2.6H2O, 1.5 g of Tris base, 10 g of casamino acids, and 5 g of yeast extract. The medium was supplemented with the following amount of supplements per liter: 0.04 g of L-cysteine monohydrochloride, 0.01 g of FeSO4.7H2O, 0.15 g glutathione (reduced), 0.4 g of L-ascorbic acid, 0.004 g of niacin, and 0.02 g of CaCl2.2H2O. The supplement was prepared in a concentrated form (100x) and filter sterilized. Cultures were incubated on an orbital shaker (170 rev/min) for 24 h at 35 °C.

This 200-mL culture was transferred to 10-L flask containing 3 L of liquid medium having the same composition as the 200-mL medium. The 3-L culture was incubated on a rotary shaker (190 rev/min) for 24 h at 35 °C. When optical density (OD) at 580 nm of the 3-L culture was more than 1.0, it was transferred into a fermenter containing 30 L of defined medium. Heptakis solution was sterilized by filtration and added into the fermenter at a concentration of 1 g/L of liquid medium. The bacteria were grown in batch mode in a fermenter of 50 L operating volume at 35 °C. The fermentation was maintained at a temperature of 35 °C, a dissolved oxygen level of 25% by cascading with an aeration of 3 L/min to 21 L/min and the agitation rate from 130 rpm to 290 rpm. The pH was controlled by a solution of glutamic acid and FeSO4.7H2O in hydrochloric acid. Foaming was controlled by adding antifoam 204 as needed. The culture was harvested after reaching an OD 7.5±0.1.

TFF Membrane and System

The TFF was carried out using a microporous membrane of pore size 0.45 μm. The construction material of the membrane was Durapore, i.e., hydrophilized polvinyledene difluoride (PVDF). The Prostak TFF module (Millipore, MA) consists of pre-assembled, pretested, and prebonded PVDF membranes on polysulfone plates and available in filtration area. This open channel module is designed to operate up to 5.6 bar, is steam sterilizable, and can be regenerated by cleaning.

An automated microfiltration (MF) system (Millipore, Bangalore, India) consisting of a 75-L jacketed tank was used in this study. The fluid to be clarified entered from the tank to a rotary lobe pump of capacity 1.8 m3/h at 4 bar pressure and 750 rpm, after the inlet pneumatic diaphragm valve was opened. The feed temperature and pressure were monitored by transmitters. Differential pressure across the installed device was controlled by throttling the motorized diaphragm valve on retentate (fluid that is retained by the membrane) line and retentate flow was measured by an electromagnetic flow meter. The permeate (fluid that passed through the membrane) line was equipped with a flow meter, a pressure and conductivity transmitter, and a peristaltic pump to control the permeate flow. The total system was steamed-in-place (SIP) and a steam trap was used at drain points to remove the condensate after SIP. The flow rate, volume of fluid at permeate and retentate line were monitored continuously during TFF pressures at the feed, permeate, and retentate line.

Clarification of Fermentation Harvest

The TFF system was installed with a 0.45-μm open channel Prostak device (PVDF membrane, Millipore). The MF filtration module consisted of four stacks with total filtration area 0.34 m2. The system, fitted with the membrane, was flushed with 20-L water for injection (WFI). Integrity of the MF membrane and installation was checked by a diffusion test using an automated integrity testing machine (IT4, Millipore). The system was then drained. SIP of the system was conducted, steam pressure inside the system was maintained at 1.2 bar and 123 °C for 30 min, and then the system was left at a positive pressure of 1 bar. The next day, it was observed that the system was at a positive pressure of 0.4 bar at 28 °C.

Thirty litres of B. pertussis culture was transferred aseptically into the system tank and clarification was carried out at constant transmembrane pressure (TMP). A secondary peristaltic pump in the permeate side was used for permeate flow control. The feed was concentrated 10 times (30 to 3 L). Temperature, pressure, and volume were monitored at the feed, permeate, and retentate line during the operation.

Assay Methodologies

The PT and FHA were quantized by ELISA with purified antigens as standards. Microtiter plates (Nunc Maxisorp, Roskilde, Denmark) were coated with either sheep anti-PT antibody (PT ELISA) at a dilution of 1:4,000 or sheep anti-FHA antibody (FHA ELISA) at dilution of 1:12,000 in 50-mM sodium carbonate buffer (pH 9.6). Plates were incubated overnight at 2–8 °C. Volumes of 0.1 mL were used in all steps and the plates were washed four times between incubations by using 100-mM phosphate-buffered saline containing 0.05% (w/v) Tween 20. The plates were blocked for 1 h at 37 °C with 2% (w/v) bovine serum albumin (BSA) in phosphate-buffered saline (PBS; pH 7.4). The plates were washed and culture supernatants were serially diluted and added to the wells, and then the plates were incubated at 37 °C for 1 h. Mouse anti–PT at dilution of 1:1,000 and mouse anti-FHA at dilution of 1:500 in 1% (w/v) BSA in phosphate-buffered saline (PBS; pH 7.4) were added to the plates. The plates were incubated for 1 h at 37 °C. Bound antigen was detected by using goat antimouse horse reddish peroxidase conjugate with a dilution of 1:7,000 (PT ELISA) and 1:10,000 (FHA-ELISA) in 1% BSA in phosphate-buffered saline. The plates were incubated for 1 h at 37 °C and were washed and developed with tetramethylbenzidine and 0.006% hydrogen peroxide in 100 mM citrate-phosphate buffer, pH 5.0. The reaction was stopped by adding 2NH2SO4. The enzyme substrate reaction that developed at 37 °C was measured at 450 nm on a Universal Microplate reader (EL-800 with KC4 data analysis software, BioTek Instruments, Inc., Vermont, MA).

Results and Discussion

There are different options for clarification and acellular pertussis antigen recovery. Centrifugation is the most commonly used tool in primary clarification. It can handle considerably high concentrations of insoluble material in the feed. A critical drawback of centrifugation is that cell disruption can occur because of shear, resulting in the generation of smaller particles, which cannot be separated by a centrifuge efficiently. The precipitation efficiency of a centrifuge decreases with increasing deposition of solid sludge in the bowl resulting in frequent discharge, which in turn, decreases product yield and increases process time. Buffer or water used for bowl flush can create osmotic differences between culture broth and flush fluid, resulting in cell lysis and further downstream complication. Dead end or normal flow filtration is unable to take high particle load and the residue on the filter is not recoverable, and hence, is not a feasible option for the clarification of B. pertussis fermentation broth, where the whole cells (containing PRN) and the filtrate containing PT and FHA are the products of interest.

In this microfiltration study, recirculation is accomplished in a closed loop (Figure 1) using a TFF module. The feed pump allows the product present in the feed tank into the TFF module tangentially across the membrane. Materials smaller than the pore size of the membrane are able to pass through the membrane (B. pertussis fermentation media containing the antigens PT and FHA), and are collected aseptically. The membrane retains the larger particles (retentate) containing B. pertussis, cells from where PRN is recovered.

Figure 1. A schematic diagram of microfiltration flow pattern

The retained components do not build up at the surface of the membrane. Instead, they are swept along by the tangential flow. This feature of TFF makes it an ideal process for finer sized-based separations.

The pump at the permeate line is used to control polarization across the membrane by offering permeate flow control and low pressure from the permeate side. This type of control ensures that the flux and TMP are low and stable across theflow channel and enables microfiltration to be carried out causing a very low shear to the cells.

Throughout the TFF experiment, pressure at the feed (P1), permeate (P2), and retentate (P3) were monitored and the volume of permeate collected per unit time (VP), retenate (Qr) was noted. The permeate flux was calculated by measuring the permeate flow rate per unit membrane area. Details of the process parameters are presented in Table 1.

Table 1. Microfiltration process parameters

The inverse relationship of flux and TMP is commonly observed in a microporous TFF operation. Higher TMP causes greater compaction of the cake layer deposition against the membrane, resulting in a higher resistance to filtrate flow rate and a lower flux. Therefore, the flux and TMP control offered by TFF system in this investigation helps to optimize the flux while also preventing cell damage because of low shear, and thereby, potentially maximizing the passage of soluble FHA and PT components through the membrane.

Table 2. Recovery of antigens

Flux and TMP were constant throughout the MF operation, with a high yield (Table 2) of cell mass and FHA, PT antigens in the permeate (results were mean value of five studies). The initial high flux was because of ramping up of the pump to appropriate feed flow rate, which was stabilized after 20 min. The flux remained steady and the average flux obtained in the experiments was 37.75 L/m2/h at an average TMP of 0.135 bar. The relationship of Flux and TMP is shown in Figure 2.

Figure 2. Flux and TMP curve during the unit operation

The optimal maintenance of cross flow (retentate flow rate) and well regulation of TMP enhanced and maintained consistent recovery in all the five experiments. The next goal was the establishment of an effective cleaning regime ensuring repeated usage of the membrane to reduce the downstream cost. Twenty liters of WFI was recirculated for 5 min and then drained off. WFI recirculation was done twice. The system was sanitized with sodium hypochlorite solution (200 ppm of chlorine), recirculated for 30 min and drained off. Normalized water permeability (NWP) was calculated before and after cleaning. After each experiment, the NWP was restored to its original value, thereby establishing the cleaning effectiveness.

The B. pertussis fermentation broth must be clarified quickly to increase the recovery of the acelullar pertussis vaccine antigenic components. Conventional methods like centrifugation are cumbersome, time consuming, and have limited scalability. Membrane-based technology in TFF mode can ensure easy and conventional scale-up with a high recovery percentage. An effective protocol suitable for scale-up is presented in this article. Because of the same channel geometry in different modules of Prostak, i.e., 2, 4, 10, and 20 stacks of effective filtration area 0.17, 0.33, 0.84, and 1.7 m2 respectively, it provides a suitable technology for manufacturing at each scale. After a clarification step, the Prostak modules can be regenerated and sanitized, providing reusability with validated steam sterilization capacity for 20 cycles. The higher recovery is mainly because of the low hold up volume of the system, the low protein binding nature of the membrane, and also because the operation takes palce under low shear condition. The Prostak module's design allows aseptic processing without cross contamination or aerosol generation.


The authors are thankful to Lalit Saxena and Gajanan Joshi of Millipore India Pvt. Ltd. for technical assistance during the study.

RAKESH KUMAR, PhD, is a director and S.V. KAPRE, PhD, is an executive director of the Serum Institute of India, Ltd., Pune, India, PRIYABRATA PATTNAIK, PhD, is a technical manager of the biomanufacturing sciences network at Millipore Singapore Pte. Ltd., Singapore, +65.6403.5308, priyabrata_pattnaik@millipore.comSUBHASIS BANERJEE, PhD, is a manager of process applications and M.S. MAHADEVAN is a vice president of the bioprocess division at Millipore India Pvt. Ltd., Bangalore, India.


1. Cherry JD, Heininger U. Pertussis and other Bordetella infections, In: Feigin RD, Cherry JD, eds. Textbook of Pediatric Infectious Diseases, 4th edn. WB Saunders; Philadelphia, PA. 1998:1423–40.

2. Conly JM, Johnston BL. The role of the acellular pertussis vaccine and the demise of the 'Pertussis Pete'. Can J Infect Dis. 2001:98:15–17.

3. Pertussis Vaccination : Use of Acellular Pertussis vaccines among infants and young children, Recommendations of the advisory committee on immunization practices (ACIP); Morbidity and mortality weekly report, US Department of Health and Human Services. 1997;22: Mar 28.

4. Corbel MJ, Corbel MJ, Toxicity and potency evaluation of pertussis vaccine. Vaccines. 2004:89–100.

5. Sundaran B, Palianiappan C, Rao YUB, Boopathy R, Bhau RLN. Tangential flow filtration technology applicable to large scale recovery of diptheria toxin. J Biosci Bioeng. 2002;94:93–98.

6. Rao YUB, Mahadevan MS, Michaels LS. Evaluation of microporous tangential-flow filtration in the production of diptheria and pertussis vaccines. Pharm Technol. 1992;16:102-110.

7. Stainer DW, Scholte MJ. A simple chemically defined medium for the production of phase I Bordetella pertussis. J Gen Microbiol. 1970;63:211–220.