3R Initiative Within Roche’s Global QC Network

In 1959, British scientists William Russel and Rex Burch published the book “The Principles of Human Experimental Technique,”1 in which they defined the 3R principles. These principles are aimed at completely avoiding (replacement) or reducing (reduction) the number of animals used for testing and improving animals’ living conditions (refine improvement). In Europe, the objectives of the 3R principles were established more than 35 years ago in a convention for the protection of vertebrates used for experimental and other scientific purposes.2

This agreement was ultimately the basis for the European Union (EU) Directive 2010/63/EU on the protection of animals used for scientific purposes.3 This directive embedded the internationally recognized 3R principles into European law for the first time. Hence, all national and European authorities (e.g., the European Pharmacopoeia, a council of the EU) are bound by these requirements, and an annual reporting has been established to monitor the progress of the implementation of the 3R principles in the EU.

The ALURES Statistical EU Database contains data on the use of animals for scientific purposes collected by the member states.4 Data extracted from the statistical reports uploaded to ALURES Statistical EU Database (see Table 1) indicate that between 1991 (the first report available on the database) and 2020, the number of in vivo assays decreased from almost 11.5 million to ~10.5 million per year. It should be noted that the number of EU member states reporting their annual data has steadily increased from 11 (in 1991) to 29 (28 member states and Norway) in 2020. It must therefore be assumed that in the years 1991–2008, when there were fewer member states reporting their data, the number of in vivo assays performed in the EU was higher than indicated in the annual reports.

In the EU, basic research and development in human and veterinary medicine are the areas that use the highest number of animals for scientific purposes. The percentage of animal experiments for regulatory use to satisfy regulatory requirements (~15%) and for routine production (~5%) has remained relatively constant.

Many pharmaceutical companies are globally active and market their products internationally. This means that approval of alternative methods to in vivo assays by regulatory authorities would be required internationally to achieve a reduction in animal consumption. As part of the harmonization of the requirements for the validation and implementation of alternative methods, alternative approaches are discussed in international forums and harmonized internationally.

These forums include groups like the Organization for Economic Co-operation and Development (OECD), the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH), the International Cooperation on Harmonisation of Technical Requirements for Registration of Veterinary Medicinal Products (VICH), and the International Cooperation on Alternative Test Methods (ICATM). The overarching aim of these groups is to promote increased international cooperation and coordination in the scientific development, validation, implementation, and regulatory use of alternative approaches.

In Vivo Assays in QC

Animal-based in vivo QC assays are typically used for the development and manufacturing of biopharmaceutical products. These include the qualification of eukaryotic cell lines, specific testing for viral clearance of products, demonstration of the absence of pyrogens to ensure patient safety, and the determination of potency of the manufactured biopharmaceuticals.

The majority of in vivo assays used in QC can be avoided without losing the appropriateness of the quality assessments. Following the 3R principles, all in vivo assays should be assessed and replaced or reduced. This would also help meet increasing regional and global sustainability goals. Novel analytical/QC technologies enable the replacement of existing in vivo assays while ensuring the efficacy of the active pharmaceutical ingredient and the absence of contamination that poses a risk to patients.

Cell Bank Testing

In most cases, complex biopharmaceutical molecules are manufactured using eukaryotic expression systems (typically using recombinant Chinese hamster ovary [CHO] cells). To guarantee a robust manufacturing process and a constant product quality over a long period of time, so-called cell banks are used.

Cell banking is the process of preserving and storing cells for future manufacturing purposes. It typically involves a two-tier system, establishing a master cell bank first and a working cell bank from a single cell source second, expanding the cells, and cryopreserving them for long-term storage. It is crucial to comprehensively characterize and test these cell banks of the recombinant cell lines to ensure identity, stability, performance, quality, and safety.

To guarantee the quality and safety of the cell banks, various tests must be carried out to demonstrate the absence of a contamination caused by “adventitious agents” in the cell bank(s). Traditionally, in vivo assays have been used to detect viral contamination in cell banks.

In Vivo Virus Detection Assay

The in vivo virus detection assay5 is used to detect viral contamination that could contaminate mammals—the CHO host cells but also patients. For this purpose, the cell bank cell culture fluid is injected into adult and suckling mice, as well as embryonated eggs. Depending on the source and nature of the manufacturing cell lines, additional animal species, (e.g., hamsters) may need to be tested. After injection, the health of the animals is monitored, and any abnormality is investigated to determine the cause. For the assessment of one cell bank using the in vivo virus detection assay, 75 mice (15 adult/60 suckling) and 80 embryonated eggs are required.

In Vivo Antibody Production Test

These in vivo assays should be used when the potential exists for exposure to viruses of a specific animal species. For example, murine viruses that are specific to certain rodent mammals can be detected using an immunologically based in vivo assay, the antibody production test. For this purpose, the test article is injected into virus-free, highly susceptible natural hosts (i.e., mice or hamsters) and the antibody level in the serum of the animals is assessed after a specified time using sensitive and specific serological assays. Examples of such tests that are carried out for the cell bank control are the mouse antibody production (MAP) test and the hamster antibody production (HAP) test.

Specific molecular biological assays—for example, nucleic acid amplification techniques (such as the polymerase chain reaction [PCR]) or next-generation sequencing (NGS, or massively parallel sequencing [MPS] or high-throughput sequencing [HTS])—can replace the in vivo assays described previously. The revised version of the ICH Guideline Q5A(R2) on viral safety evaluation of biotechnology products derived from cell lines of human or animal origin5 encourages applicants to use these alternative assays.

Roche has recently implemented a multiplexed degenerated PCR6 into their control concept that combines the advantages of PCR, speed and sensitivity, with a broad specificity to detect a large number of virus variants. As a result, the species-specific viruses that can be detected by the in vivo MAP and HAP tests can be detected by a PCR-based in vitro assay. In recent years, the introduction of the PCR-based virus detection method reduced the need for animals by up to 90 mice and almost 45 hamsters per year.

Pyrogen Testing

Pyrogens are fever-inducing substances of various origins, like bacteria (dead or viable), fungi, viruses, or even chemicals. The innate human immune system reacts to pyrogens by releasing cytokines, which lead to an increase in body temperature (fever) and inflammation. Severe (over)reaction of the innate immune system (adverse reaction) can lead to an unregulated release of cytokines and ultimately to shock, multiple organ failure, and even death. For this reason, only therapeutics that are free of pyrogens may be marketed and health authorities require testing for pyrogens at various points in the production process. Three different tests are carried out for this purpose:

Rabbit Pyrogen Test

Owing to a comparable pyrogen sensitivity in humans and rabbits, the implementation of the in vivo rabbit pyrogen test (RPT) in the 1940s was able to significantly increase the safety of pharmaceutical products. The test is based on the reaction of a rabbit’s innate immune system, which, like humans, reacts to pyrogenic substances by increasing body temperature and can therefore detect the majority of pyrogens. This includes endotoxin derived from gram-negative bacteria as well as non-endotoxin pyrogens derived from gram-positive bacteria, fungi, virus, or chemicals.

In the RPT, the product being tested is injected intravenously into the ear vein of rabbits. The temperature of the rabbits is monitored over a defined period of time. Finally, either the sum of the body temperature increase is calculated7 or the increase of the body temperature of individual rabbits8 is assessed. If the summed response does not exceed a threshold value (European Pharmacopeia acceptance criterion) or no rabbit shows an individual rise in temperature of 0.5°C or more above its respective control temperature (US Pharmacopeia acceptance criterion), the product meets the requirements for the absence of pyrogens.

The RPT currently consumes a significant number of animals. In the EU, ~50,000 animals were used for this test in 2015 and ~25,000 in 2021.4 The assumption for worldwide use is ~400,000 rabbits per year. In response to animal welfare objectives, multinational measures have been initiated in the EU to avoid the RPT. One notable initiative is led by the European Directorate for the Quality of Medicines & HealthCare, which aims to remove the RPT from the European Pharmacopeia.9^, 10^, 11^,12 This objective involves creating new or revising existing European Pharmacopeia General Chapters and the revision of around 60 monographs to replace the RPT with either the monocyte activation test (MAT) or the bacterial endotoxins test (BET) (described as follows).

Monocyte Activation Test

The monocyte activation test (MAT)13 is an in vitro assay that mimics the reaction of the human innate immune system to pyrogens. A cell suspension containing monocytic cells (either peripheral blood mononuclear cells [PBMCs] or whole blood) produces cytokines (e.g., interleukins such as IL-1β or IL 6) by activation with pyrogens, and the released cytokines can be detected by an immunological assay (e.g., enzyme-linked immunosorbent assay [ELISA]).

Bacterial Endotoxins Test

Bacterial endotoxins, a subset of pyrogens, are lipopolysaccharides (LPS), a component of the outer cell wall of gram-negative bacteria. Endotoxin is a heat-stable substance and is known to be the most potent pyrogen. Because of these properties and their ubiquitous occurrence, the absence of endotoxin in the final product is of particular importance for pharmaceutical production.

In contrast to the MAT or the RPT, the BET is highly specific for endotoxins. Non-endotoxin pyrogens such as components of gram-positive bacteria, nucleic acids, proteins, or chemical components cannot be detected by the BET. Two different versions of the BET are accepted by the health authorities: the Limulus amoebocyte lysate–based BET (LAL assay) and the recombinant Factor C-based BET (rFC assay).

LAL assay

For almost 50 years, the LAL-based BET14^, 15 has been used for the determination of endotoxins, which rely on the blood from ancient and endangered horseshoe crabs (Limulus polyphemus or Tachypleus tridentatus). This test is based on the amoebocytes in the horseshoe crab’s “blood,” which provide the animals’ natural defense mechanism against bacterial endotoxins and other pathogens. When the amoebocytes come into contact with bacterial endotoxins, an enzyme cascade is activated, which finally leads to clot formation of the LAL. Different variants of the LAL-based BET were developed: the gel clot assay and end-point or kinetic versions of a turbidimetric and a chromogenic assay.

rFC assay

The basis for this BET16 is a synthetic rFC, the first enzyme in the LAL enzyme cascade. The activation of rFC by endotoxins leads to cleavage of a synthetic fluorogenic substrate present in the test system and thus to a fluorescent end product that can be quantified. The advantage of the rFC-based BET is that no animal-sourced reagent of an endangered species, such as the horseshoe crab, is required.

Avoiding In Vivo Pyrogen Tests

To avoid in vivo assays or the use of animal-sourced reagents for pyrogen testing, we have initiated several projects, which are discussed in the following sections.

Replacing the in vivo RPT with the BET

An analytical variation was submitted to replace the in vivo RPT assay with the LAL-based BET. It is one of our legacy products for which the RPT was originally filed many years ago as the pyrogen test method to release produced batches. The rationales for replacing the in vivo RPT were a) improved microbial control of the manufacturing process to prevent any unintentional introduction of pyrogens into the manufacturing process and b) prior knowledge based on hundreds of batches produced and released that demonstrated that the product itself was not pyrogenic. The analytical variation for lot release testing was accepted by the US Food and Drug Administration (FDA).17 This update to the control system means that beginning in 2023 (the year of approval of the analytical variation) approximately 40 rabbits per year no longer need to be used for the release-relevant in vivo RPT.

Replacing the in vivo RPT with the MAT

During the development of a new manufacturing process for biologics, it must be shown that the manufactured biotherapeutics are free of pyrogens. In Europe, the MAT13 can be used to demonstrate the absence of pyrogens. Unfortunately, in most counties outside of the EU, the MAT is still considered as an “alternative method” and in this respect the absence of non-endotoxin pyrogens must be demonstrated using the in vivo RPT.

For the US, this requirement is described in a federal law, 21 CFR 610.13(b).18 However, according to 21 CFR 610.9,19 it is acceptable for the in vivo RPT to be waived if an alternative method is used that is equivalent to the in vivo RPT. The equivalence is usually demonstrated by analyzing a few representative batches in parallel with the in vivo RPT and the alternative method—typically the LAL-based BET. If the LAL-based BET is negative (demonstration that no endotoxins are present in the product) and the in vivo RPT is negative for the same batches (demonstration that neither endo-toxins nor non-endotoxin pyrogens are present in the product), this is considered proof of equivalency. This means batches produced in the future may be released on the result of the LAL-based BET.

Following the described approach, it is required that some manufactured lots must be tested in parallel with the BET and the in vivo RPT. This concept is therefore still based on in vivo assays. Roche recently submitted a dossier for a pilot product, firstly describing a validated MAT and secondly demonstrating the absence of pyrogens using the in vitro MAT. This revised validation strategy, which makes the in vivo RPT superfluous, was recently approved by the US FDA20 and is to be used in the future for all product and process developments. With the revised validation concept, at least 9–24 rabbits can be saved for each product under development.

Replacing the LAL-based BET with the rFC-based BET

Pharmaceutical products for parenteral administration must not be contaminated with endotoxins during the production process. As previously described, for almost 50 years, the LAL-based BET has been used for the determination of endotoxins, which rely on the blood from ancient and endangered horseshoe crabs (Limulus polyphemus or Tachypleus tridentatus).

Only a component of the “blood” of the horseshoe crabs is used to produce the LAL, so the LAL-based BET is by definition not an in vivo assay. However, a replacement of the LAL-based BET is desirable for ethical reasons. This is because the horseshoe crabs must be removed from their natural habitat for blood collection, and not all horseshoe crabs survive when the “blood” is taken from them. The solution to reduce the impact of horseshoe crab dependency is the use of the rFC-based BET.16

The majority of in vivo assays used in QC can be avoided without losing the appropriateness of the quality assessments.

As part of our 3R initiative, the LAL-based BET is currently being replaced by the rFC-based BET for testing water samples and product samples. For more than a year, water testing using the rFC-based BET has been successfully carried out at a pilot site in our QC network. At this site alone, more than 11,000 LAL-based BETs can be replaced with the recombinant alternative. The rFC-based testing of water samples is currently being transferred to other internal QC sites.

In a parallel project, the testing of product samples using the rFC-based BET is also being evaluated. Recently, the dossier of a pilot product was sub-mitted worldwide, in which both the in-process control samples and the release-relevant samples are tested using the rFC-based BET. This analytical variation has already been accepted by some health authorities (e.g., FDA21 and European Medicines Agencies/Committee for Medicinal Products for Human Use [EMA/CHMP]). In the next step, the method will be implemented worldwide in our QC network and will be used for future products.

Potency Testing

Potency testing is crucial for assessing therapeutic effectiveness and biological activity as well as ensuring consistent batch-to-batch quality of pharmaceutical products. Common approaches include various product-specific types of bioassays, such as cell-based, receptor-binding, and neutralization assays, which directly assess the product’s potency to stimulate or inhibit specific cell growth. In addition, in vivo assays are still used for longer marketed products. To meet 3R requirements, replacing the in vivo potency assays should be considered, if possible, with in vitro cell-based assays.

For one recombinant biopharmaceutical product, thousands of mice are used per year to demonstrate the biological activity of the produced batch-es via an in vivo potency assay. To avoid these in vivo assays in the future, we developed an in vitro reporter gene assay to determine the bioactivity. For this purpose, a genetically modified cell line is used to mimic the mode of action in vitro. The cell line enables the assessment of the pathway from the binding of the ligand to its specific receptor until the activation of gene expression.

In the in vitro reporter gene assay, the gene that is finally activated is modified. In the new in vitro reporter gene potency assay, it is not the formation of new cells that is achieved, but the expression of an enzyme. Ultimately, the activity of the enzyme can be used for an alternative readout, e.g., bioluminescence, ELISA, or another suitable readout system.

We recently obtained approval from the US FDA22 to use the in vitro reporter gene assay for the determination of the biological potency of the produced biopharmaceutical product. With the implementation of this analytical variation, the use of several thousand mice per year that are currently required for the in vivo potency assay can be avoided.

Conclusion

The sustainability goals of numerous regulatory authorities worldwide have initiated the development of alternative methods as a replacement for in vivo assays. A thorough method validation was performed for each of the alternative methods. As part of the validation, the relevant parameters23 of the alternative methods were investigated. In addition, the method validations demonstrated that the alternative methods are comparable (non-inferior) to the traditional in vivo assays used to date and provide equivalent results. They also ultimately “provide assurances of the safety, purity, potency, and effectiveness of the biological product equal to or greater than the assurances provided by the method or process specified in the general standards or additional standards for the biological product.”19 

The submitted results of the method validations of the alternative methods were reviewed by the competent authorities and as an outcome of this review, the use of the alternative methods were approved as a replacement for the traditionally used in vivo assays.

The projects described demonstrate the potential that 3R initiatives have in the strongly regulated quality and manufacturing environment and show that it is possible to reduce dependence on in vivo assays. By using alternative methods and considering the 3R principles in the control concepts of biopharmaceutical products, the industry can contribute to significantly improving animal welfare.

Acknowledgements

The author would like to thank the responsible project leaders and the numerous project members of the 3R projects described in this article for their efforts and commitment. Unfortunately, it is not possible to individually list and thank all Roche colleagues who are working on the replacement of in vivo assays by alternative methods for their many years of tireless commitment.