Elsevier

Vaccine

Volume 25, Issue 16, 20 April 2007, Pages 2994-3000
Vaccine

PAT for vaccines: The first stage of PAT implementation for development of a well-defined whole-cell vaccine against whooping cough disease

https://doi.org/10.1016/j.vaccine.2007.01.015Get rights and content

Abstract

Since variation in process time and process output is commonly accepted to be inevitable for biological processes, application of Process Analytical Technologies (PAT) on these processes is challenging. In this paper the applicability of PAT on the cultivation of Bordetella pertussis bacteria as part of the manufacture of a vaccine against whooping cough disease is investigated. Scrutinizing and eliminating the most prominent sources of variance make the cultivation process step highly reproducible. Furthermore, the use of DNA microarrays allows investigation of how disturbances influence cellular physiology and product quality. Marker genes for product quality were identified, providing the means to quantitatively assess product quality, which is hardly possible using the mandatory animal tests for product quality. The tools and results described in this paper, combined with suitable on line measurements, can make full PAT application for this process step possible. Ultimately, the process can be designed and controlled towards consistent end product quality.

Introduction

Production processes for biologicals are generally considered less controllable and more susceptible to variation than processes for chemical drugs due to poor process understanding. Vaccines are considered to be at the low end of this spectrum. The production of vaccines has historically only been considered viable as a low-cost, low-tech industry and this became even more difficult after the introduction of GMP in 1980. Despite all the breakthroughs in immunological sciences of the past decades, the production processes for vaccines are essentially unchanged. Only recently vaccines have attracted the interest of large pharmaceutical companies for being potential block-buster candidates [1]. This means that vaccines are more and more becoming high-value biologicals. This generates a drive for more robust and efficient manufacturing processes for vaccines.

The US Food and Drug Administration (FDA) has recognized the need for robust and efficient manufacturing throughout the pharmaceutical industry. The high regulatory burden in this industry prevents continuous process optimization and process adaptations as a means for increasing manufacturing efficiency. In order to overcome these regulatory barriers, the FDA has launched the Process Analytical Technology (PAT) Initiative in 2003 [2].

PAT is considered by the FDA to be the desired future state of pharmaceutical manufacturing. It requires a risk-based approach in which the critical process parameters or attributes are assessed for their risk for product quality. Monitoring and control of these parameters can assure good process performance and even the quality of the product as it is formed [3]. This requires an on line monitoring and control system that can measure the critical parameters during processing and take appropriate action upon any deviation. Although the application of PAT in a vaccine manufacturing process may seem challenging, a science-based approach has much to win in this industry, since production processes for vaccines are usually highly empirical. Before a PAT system can be designed, the process needs to be sufficiently understood and the critical process attributes that need to be monitored and preferably be controlled on line have to be identified. In this paper these first stages of PAT application are investigated for the cultivation process step of a vaccine against whooping cough disease.

For inactivated or attenuated live vaccines product quality is primarily determined during cultivation of the actual pathogen. For acellular or subunit vaccines this might be the cultivation of the host cells (i.e. CHO cells or E. coli bacteria), which produce the required vaccine components. The focus of this work is on a whole-cell vaccine against whooping cough disease. This vaccine is based on a heat-inactivated suspension of the pathogen, the Bordetella pertussis bacterium. Especially with such an inactivated whole-cell vaccine, the cultivation process step is the most relevant process step for PAT application, because in this step product quality is mainly determined.

A whole-cell vaccine might in principal provide a broader immune response than an acellular vaccine, because it inherently contains all necessary components for inducing protection. Several studies have indicated that whole-cell vaccines may have a better efficacy than acellular vaccines [4], [5], [6], but the side effects after vaccination have resulted in a strong decrease of their application. These side effects (i.e. swelling, fever) are mainly caused by lipopolysaccharide (LPS). The presence of LPS anchored in the outer membrane is relatively constant under normal production conditions [7], whereas the expression of the outer-membrane proteins that are important for inducing a protective immune response can vary depending on cultivation conditions [8]. Optimal outer-membrane protein composition can therefore be considered a critical attribute for product quality. More precisely, the expression levels of virulence factors are critical for product performance.

The mechanism through which B. pertussis regulates the expression of virulence factors has been extensively studied [8], [9], [10]. Most known virulence factors are under control of a two-component regulatory system named BvgA/S. This system functions as a molecular switch that acts to activate or repress virulence factor expression depending on environmental conditions (Fig. 1). In order to obtain a high-quality product, this switch needs to be in the bvg+ mode, resulting in bacteria that express virulence factors abundantly.

Having defined the properties of the product that are critical for product quality, accurate measurement of these properties is the next step for PAT application. This requires much more precise information than is provided by the mandatory animal tests that are currently used for this vaccine [11]. DNA microarrays for analysis of gene expression levels and profiles are proposed for assessment of product quality at the end of cultivation. mRNA expression levels are expected to serve as an indicator for the expression of virulence factors and other proteins relevant for prediction of product quality and process performance.

The availability of the full genome sequence of B. pertussis[12], created the opportunity to design a DNA microarray covering 93% of the B. pertussis open reading frames. Recent research has shown that the use of DNA microarrays can accurately determine the virulence state of the bacterium [13], [14]. This technique can be applied in several manners. First, the genes crucial for virulence, and thus product quality, can be assessed. Second, overall expression levels at the end of cultivation can be compared with other batch runs. Good correlation between overall mRNA expression levels will indicate comparable end-product quality, whereas disturbances in expression levels can be identified and investigated.

After the desired product is defined and its quality can be accurately assessed, the biggest challenge that remains is the reproducible performance of the actual manufacturing process. The ability to do this cultivation in a robust and reproducible manner is a prerequisite for implementation of PAT, as it demonstrates process understanding and the ability to control critical process variance. Bacterial cultivation has many sources of variance, some of which could risk end quality to be out of specifications. Assessment of all relevant sources of variance and subsequent assessment of their potential risk for end-product quality, although complicated, could be one of the most rewarding exercises in the PAT trajectory, because improvement of process reproducibility is of key importance for robust processing.

In this paper we demonstrate the first stages of PAT application in the cultivation process of B. pertussis bacteria for a vaccine against whooping cough disease. First, the elimination of the most obvious sources of variation reduces the variation of the cultivation process. Next the use of DNA microarrays for identification of the relevant genes affecting product quality is demonstrated. Finally, we show how DNA arrays can be used to check the reproducibility of the cultivation with respect to the entire transcriptome and the relevant marker genes.

Section snippets

Bacterial strain, media and growth conditions

The B. pertussis 509 strain (Netherlands Vaccine Institute, Bilthoven, The Netherlands) used in this study was collected in 1963 as a clinical isolate and used as vaccine strain in the combined diphtheria, pertussis, tetanus and polio (DPTPolio) vaccine until 2005.

All bacterial cultures were grown in chemically defined THIJS-medium [15], which consists of basic medium and supplement. Supplement (1%, v/v) was added to the basic medium shortly before inoculation. The pH after addition of

Bacterial growth

The cultivation method using THIJS medium described above resulted in a highly reproducible cultivation. Comparison of 11 bioreactor runs shows a high correlation between runs (R2 = 0.99) and low variance (S.E. = 0.055) (Fig. 1).

Genes involved in virulence

Chemical induction of an avirulent state, using 50 mM MgSO4 or 20 mM niacin, and subsequent comparison with normal cultivation conditions revealed the genes involved in virulence of the B. pertussis 509 strain. These genes were compared with the core regulon proposed for the

Discussion

Variation in process performance is a high risk for inconsistency of product quality. Identification of the factors that cause process variation will enable their control, which in turn will result in more robust and consistent processes. Biological processes such as the cultivation of bacteria are usually considered variable and hard to control. If, however, a bacterium is given only a limited number of degrees of freedom, it will adopt a predictable growth pattern, as demonstrated here. A

Conclusions

Generally, the cultivation of vaccines is considered to have high inherent variance and therefore not as suitable for PAT application as chemical (synthesis) or physical (blending) process steps. This paper demonstrates that elimination of the major sources of variance leaves a cultivation process step that is highly reproducible. Application of modern techniques such as DNA microarrays provides a tool for science-based evaluation of product quality. Disturbances in mRNA expression levels of

Acknowledgements

The authors gratefully acknowledge Dr. J. Pennings for his help with the microarray analysis and data processing, Dr. B. Zomer for NMR measurements and Dr. M.J. Vansteensel for help with the artwork. This work was partly funded by Senter Novem, The Netherlands, Grant Number TGSE3067.

Reference (18)

There are more references available in the full text version of this article.

Cited by (32)

  • Microbial carbohydrate-binding toxins – From etiology to biotechnological application

    2022, Biotechnology Advances
    Citation Excerpt :

    The inhibition of PHB production provided a 7-fold increase in biomass formation (6,5 g/L cell dry mass) and consequently an overall volumetric PTx yield of 12 mg/L. The experimental findings acquired from PTx production on a large-scale were implemented in B. pertussis whole-cell production on a large-scale (Streefland et al., 2007; Thalen et al., 2008). STx-1B is one of the most studied carbohydrate-binding toxins characterized by successful scale-up in heterologous systems.

  • Headspace solid-phase microextraction (HS-SPME) combined with GC–MS as a process analytical technology (PAT) tool for monitoring the cultivation of C. tetani

    2018, Journal of Chromatography B: Analytical Technologies in the Biomedical and Life Sciences
    Citation Excerpt :

    Thus, vaccines as vital biological products, require highly robust and efficient manufacturing methods for their production [1]. In 2003, the US Food and Drug Administration (FDA) introduced Process Analytical Technology (PAT) as a reliable tool in the pharmaceutical industry for increasing manufacturing efficiency and improving product quality [2,3]. In fact, FDA has affirmed PAT as the desired future state of the pharmaceutical industry.

  • Comparative analysis of different transformed Saccharomyces cerevisiae strains based on high-throughput Fourier transform infrared spectroscopy

    2017, Journal of Biotechnology
    Citation Excerpt :

    To further optimize established processes, and to efficiently control large-scale productions, biotechnological companies are encouraged to increase the knowledge of the cultivation process based on the development of new analytical tools enabling the acquisition of a significant amount of data from different experimental conditions. This knowledge will lead to faster and more efficient process development related to scale-up steps and, consequently, will decrease the overall development costs and enable efficient control procedures (Hakemeyer et al., 2012; Mercier et al., 2013; Streefland et al., 2007). However, for increasing the bioprocess understanding and speed up the development process, FT-MIR spectroscopy (650–4000 cm−1) may present advantages once this spectral region covers the fundamental vibrations of most common chemical bonds in biomolecules.

  • Method for the simultaneous assay of the different poliovirus types using surface plasmon resonance technology

    2013, Vaccine
    Citation Excerpt :

    From a regulatory perspective, Process Analytical Technology (PAT) is an important initiative that can be applied to many types of vaccine processes [24]. For example, PAT has been successfully used to increase the robustness of the production protocol for the whole-cell whooping cough vaccine [25,26]. The method reported has been created in the framework of the PAT initiative and responds to major technical advancements in the field of molecular recognition over the last decades [27].

View all citing articles on Scopus
View full text