In this section

Articles

Archived articles

PDF version, 489KB

Abstract

Botulinum toxins - some of the most poisonous naturally occurring substances - are proteins produced by the bacterium Clostridium botulinum. They are neurotoxic (i.e. toxic to nerve cells) and are responsible for causing botulism, which is most often associated with eating food containing the toxin. The toxins cause respiratory and muscular paralysis, and even death, by blocking nerve function. Although botulinum toxin A is highly toxic, it can be administered safely in extremely small doses to treat painful muscle spasms and involuntary eye muscle contractions (blepharospasm), and has been produced commercially for these and other medical purposes. It is also increasingly used for cosmetic purposes.

Every batch of a therapeutic preparation of botulinum toxin is required to be tested for potency and stability, because the toxin is a biological product with inherent variability, and these tests are carried out at several stages of the toxin production process. The mouse lethal dose 50 (LD50) assay is usually used for this purpose and involves injecting mice with the toxin, to determine the dose which will kill half of the test animals at a defined time-point. Testing at the final lot stage requires high precision, so large numbers of animals are used. The increased use of botulinum toxins for medical and cosmetic purposes has lead to increased testing and hence, increased animal use. LD50 assays are severe, and are therefore at the front line for replacement with more humane methods.

Several alternative test methods applying the 3Rs, have been developed and validated at the National Institute for Biological Standards and Control (NIBSC), and were reviewed at a recent workshop on Alternative Methods to replace the Mouse LD50 Assay for Botulinum Toxin Potency Testing. The workshop concluded that there is potential for reduction, refinement and replacement of the use of animals in botulinum toxin potency testing. Several alternative methods are now also included in the relevant European Pharmacopoeia monograph and once validated for a particular product, should be accepted and applied.

Figure 1. Clostridium botulinum bacteria as visualised under a microscope. (Image credit: Centers for Disease Control and Prevention's Public Health Image Library).

Botulinum toxin background

Botulinum toxins, proteins produced by the bacterium Clostridium botulinum , are some of the most poisonous naturally occurring substances and are responsible for causing botulism, which is most often associated with eating food containing the toxin (see Figure 1). The toxins (types A-G) act by attaching themselves to nerve endings, blocking nerve function by preventing release of acetylcholine, the neurotransmitter responsible for triggering muscle contractions, thereby causing muscle paralysis and even death. Although botulinum toxin A (BoNT/A) is highly neurotoxic, it can be administered therapeutically in extremely small doses to treat painful muscle spasms and is increasingly frequently used as a cosmetic treatment.

Commercial production and therapeutic applications of botulinum toxin have increased steadily over the last 20 years following the FDA's approval in 1989 of Oculinum, now called Botox (Allergan Inc.). For example, in 1995, the UK MHRA approved Dysport for treating strabismus a condition in which the eyes are not properly aligned with each other producing a squint and blepharospasm, an abnormal tic or twitch of the eyelid. Therapeutic applications, and the conditions for which these bacterial neurotoxins are used, continue to increase. Licenced applications now include using botulinum toxin to treat cerebral palsy in children, limb muscle paralysis after stroke, uncontrollable muscle spasms of the face and neck, facial (glabellar) lines and excessive underarm sweating (see Table 1). Unlicenced clinical applications being investigated include treating urological problems such as spastic bladder or bladder overactivity, writer's cramp, tennis elbow and wound healing after surgery (see Table 2 for a more extensive list).

Table 1. Licenced botulinum toxin products and some approved indications.

Table 2. Other (currently unlicenced) indications using botulinum toxin.

Assaying potency

As is the case with other biological products used in human medicine, the activity of the toxin must be determined for every new batch and assayed using suitable systems during the production process. Information on activity and stability is required for the concentrated active toxin, and again at the final stage of the production process. A high-precision assay is required, particularly at the final lot stage, to confirm the amount of active substance in the product before it is marketed.

The median lethal dose (LD50) assay

Due to its high sensitivity, the mouse LD50 assay was adopted by all manufacturers as a way of expressing product potency. The assay involves preparing a range of dilutions of the test toxin sample and injecting a small volume (e.g. 1.0 ml) of each dilution into the peritoneum (body cavity) of mice. The dose that kills half of the test animals is termed the median lethal dose or LD50. The potency of the toxin is expressed in terms of the LD50, where one unit is equivalent to the median lethal dose. The lower the LD50 value the more potent the toxin. To accurately determine the LD50 value of a batch of toxin, between eight and 24 animals are used per dose with up to six doses required, and up to four replicates (1-3). Doses typically fall within the range about 0.6-2.5 mouse LD50 units (i.e. 6-25 LD50/ml).

Determining potency using the LD50 test is, however, not a standardised approach. Due to small differences in the method of testing between manufacturers, such as the strain of mice or the choice of dilution buffer, different potency (LD50 values) can be determined for an identical test sample. The consequence of this is that the units used in labelling the toxin potency are product-specific and non-interchangeable (1,2). This poses problems in product use, and has potential safety consequences as new licenced, look-alike and counterfeit products enter the global market (4).

Although the mouse LD50 assay is recommended for use by national and international regulatory authorities as the primary assay method for use prior to marketing, a number of alternative methods and approaches which minimise the distress caused to test animals, use more humane endpoints, or which could replace their use, have been developed and adopted.

Measuring relative potency

As the toxin is a biological substance, the potency of the toxin will vary with each batch as well as with the assay conditions. One way to control this variability is to use a toxin with a known potency as a reference standard in a bioassay, meaning that the potency of the new batch of toxin can be defined relative to that of the reference standard. Measuring relative potency in the assay is more precise and more reproducible compared with the conventional LD50 test, and minimises any variability in the assay conditions (1,3). The LD50 test is only used to calibrate the reference standard and is not essential in the potency assay, so this approach provides opportunities for developing alternative test methods and reducing the numbers of animals currently used.

Refining potency assays: the local flaccid paralysis assay

One assay that has been proposed as a refinement to the mouse LD50 test is the mouse flaccid paralysis assay, also known as the mouse abdominal ptosis assay (6,7). This assay does not use lethality as an endpoint, but instead, it relies on a more humane endpoint observing local paralysis. The flaccid paralysis assay uses relative toxin potency to measure the dose that produces local paralysis, and is 10-fold more sensitive than the LD50 test. In the assay, a sub-lethal dose of toxin is injected subcutaneously into the top of the left hindleg (the inguinocrural region of the lower abdomen) of a mouse. The maximum dose injected is 0.2 mouse LD50 units, and the magnitude of the paralysis is directly correlated to the toxin dose administered using a scoring system. The animals are scored according to the size of the local abdominal bulge at 24 and 48 hours using a five-point scale (see Figure 2). A dose response is obtained by plotting the scores against the log toxin dose. Each test is performed in parallel with a reference preparation of known activity, which is diluted and prepared in the same way as the test sample. Estimates of relative potency are obtained using the parallel line method (see Glossary). The injected mice normally do not exhibit any signs of pain or distress, but mice given the highest toxin dose can sometimes lose about 4% of their body weight.

Figure 2. The mouse flaccid paralysis assay. The mouse on the left has been injected with botulinum toxin to assay the potency of a test sample, and shows an abdominal bulge at the site of injection, compared with the uninjected control mouse on the right.

The flaccid paralysis assay requires no specialised equipment, is relatively easy to conduct, and is rapid, yielding results in 24 to 48 hours compared with 72 to 96 hours for a typical LD50 assay. The paralysis model measures the functional activity of the toxin protein, and results show excellent agreement with the LD50 values for different products (6). Furthermore, the paralysis endpoint evaluates localized muscle effects, rather than systemic toxicity making it more relevant to the clinical use of the toxin than the LD50 assay. The assay has been validated at NIBSC, where it was found that the required toxin potency information could be determined using only 20% of the total number of animals used in the LD50 assays, therefore greatly reducing the number of animals involved in testing.

Requirements for alternative potency assays

Methods that are suitable for replacing the LD50 assay for testing the potency of therapeutic products need to fulfil specific requirements and should be assessed for their relevance and suitability for purpose when undergoing validation. For example, the replacement assay for potency testing must:

• accurately quantify small amounts (nanograms) of active toxin in the finished product, which also contains a high concentration of bulking or stabilising proteins
• provide high sensitivity, with a limit of detection close to that of the mouse LD50 assay
• be easy to use, reproducible and sufficiently robust
• be transferable for quality control in a laboratory setting
• take into account the knowledge of the mode of action of the toxin, as well as the production process of the therapeutic products.

The European Pharmacopoeia monograph on BoNT/A for injection supports the use of alternative methods subject to validation (5). A 2006 workshop on 'Alternative Methods to replace the Mouse LD50 Assay for Botulinum Toxin Potency Testing' organized by ICCVAM/NICEATM/ECVAM (see Glossary; 8), concluded that validating a refined in vivo local paralysis assay provides a realistic strategy for replacing the LD50 assay. The workshop also provided an overview of current alternative methods and approaches that, if validated, accepted and applied, could replace the use of animals, minimise the pain and distress associated with current methods, or reduce the number of animals used in assays. Four different approaches were considered to be relevant:

1. Using reference standards to measure relative potency
2. Using earlier non-lethal humane endpoints for the current in vivo botulinum toxin assay
3. Using alternative in vivo models to measure botulinum toxin activity incorporating non-lethal humane endpoints
4. Using ex vivo test models prepared from euthanized animals.

Alternative potency assays

i. Tissue-based assays

Potency assays using ex vivo nerve or muscle tissue can provide replacement alternatives to the LD50 assay, since mice or rats are only needed for donating tissue (9,10). This approach tests the effect of the botulinum toxin on nerve or muscle tissue by applying an electrical current to the tissue which is in a solution, and measuring the "twitch" response using the magnitude of the electrical reading (see Figure 3). A second reading is taken after the toxin is added to the tissue in the solution. The potency of the toxin is indicated by the decrease in the twitch response in the presence of the toxin (i.e. paralysis of the tissue), compared with the initial reading in the absence of toxin. A range of toxin dilutions is used. Time to paralysis is plotted against a pre-determined standard curve with known concentrations in LD50 units, from which the activity of the test sample can be calculated. The usual endpoint of the assay is the time taken for the magnitude of the electrical reading to decrease by half. The number of animals used to determine toxin activity is greatly reduced compared with the LD50 test (four mice per dilution compared with between eight and 24 mice per dilution; 10). Whilst this method is useful for determining toxin activity, it may have limited applicability in a quality control laboratory. These ex vivo models for potency assays require laboratory personnel trained to use sophisticated and expensive equipment, but can provide results within 2 hours. In addition, the experimental conditions can easily be varied. However, applying the toxin directly to tissue preparations in solution is far removed from intramuscular injection in vivo, and there may be poor correlation with the data from the LD50 assay or in vivo paralysis models. Additional validation studies are therefore needed to establish the usefulness and limitations of the ex vivo models, and their acceptance for batch release testing may not be straightforward. At present these models remain useful only as research tools.

Figure 3. Apparatus used to apply an electrical current to mouse or rat nerve or muscle tissue. Toxin potency is measured by the decrease in the "twitch" response (i.e. paralysis of the tissue) after electrical stimulation in the presence of botulinum toxin.

ii. Cell-based assays

Primary neuronal cells or cell lines of a neuronal lineage from rodents have been successfully used as potential replacement method for botulinum toxin potency testing (see Figure 4). In a controlled laboratory setting, cell-based systems could minimise variability between assays. However, several important challenges exist with cell-based potency assays. For example, extensive cloning or modification of the cell line may be required, which would make it more difficult to create a stable cell line. In addition, the concentration of BoNT in the final formulation is very low (in the picomolar range), and the toxin often adheres to the surface of the culture dish (surface adsorption) leaving less toxin available for uptake by the cells, which can markedly affect the results. Inactive substances used as a "carrier" for the active ingredient (formulation excipients) can also interfere with the analytical methods employed. The sensitivity of the assay using specific neuronal cell lines is directly proportional to toxin uptake by the cells. However, none of the cell line assays provide sensitivity equivalent to that of the in vivo mouse bioassays (LD50 or flaccid paralysis) and therefore, at present, they remain of limited applicability for use as a routine batch release assay for therapeutic products. Studies at NIBSC are currently focusing on primary cells such as rat spinal cord cells and differentiated cell lines of human neuronal origin with a suitable range of read outs including detecting intracellular target proteins (11). These cell-based systems may provide relevant and sufficiently sensitive markers of toxicity, whilst reducing animal use.

Figure 4. Rat spinal cord cells used in cell-based assays for testing botulinum toxin potency.

iii. Enzyme assays

Botulinum toxins consist of a light chain (L-chain), a heavy chain (Hc) receptor binding domain and a translocation domain (see Figure 5). The heavy chain targets the toxin to the axon whilst the L-chain of toxin type A is taken into the cytoplasm of the nerve cell body, where its enzyme (endopeptidase) activity enables it to specifically degrade the SNAP-25 protein required for the release of neurotransmitters from the axon endings. This toxin L-chain endopeptidase activity can be used to estimate toxin potency in a laboratory assay. The assays can detect toxin enzymatic activity in a variety of ways and can provide sensitivity comparable to that of the mouse bioassay.

One such method to assess toxin potency by detecting toxin enzyme activity has been validated at NIBSC and is now in use (12-16). The assay is precise and reproducible, and has provided potency estimates which are highly comparable with those obtained by the manufacturers using the LD50 assay, and as a result, animal use in potency testing has been greatly reduced.

Endopeptidase assays only measure toxin L-chain activity, and changes to other toxin domains, are not detected. A limitation of these assays is that the toxin potency estimates may not always exactly correlate with LD50 data. However, following validation, the assays could be employed in specific circumstances or in a tiered-testing strategy to reduce the number of mice used in current protocols. In particular such methods are able to provide valuable information on product consistency. The need to maintain a supply of critical reagents is a potential disadvantage and validation studies will need to be performed to address the question of robustness and transferability.

Figure 5. A schematic representation of the domains of the botulinum toxin. The binding domain is represented in light green, the translocation domain in dark green, and the endopeptidase domain in purple.

Summary of 3Rs strategies for botulinum toxin product potency testing

Reduction

Reducing animal use in botulinum toxin testing is possible by assaying toxin potency relative to a reference standard, rather than using the LD50 assay, meaning that fewer animals need to be used in order to provide a result of equivalent precision. A reduction in animal use could also be achieved by optimizing potency assay design and analysis and extending the shelf-lives of reference standards.

Refinement

Refining current procedures is possible by using earlier non-lethal endpoints or a non-systemic endpoint, such as local paralysis, for example. These approaches require fully functional toxin and the expression of potency relative to a reference standard. Refined methods such as the flaccid paralysis assay could be validated for use by manufacturers where fully functional assays are required and would be used instead of the LD50 assay for re-calibrating reference standards, defining activity in new bulk active toxin preparations, and validating replacement methods.

Replacement

Alternative assay methods which replace animal use are available and include ex vivo and in vitro enzyme activity models. Whilst additional validation studies are needed to fully characterize the usefulness and limitations of the ex vivo models, in vitro enzyme activity models are most suited for use as indicators of consistency, and could also reduce the number of animals used if applied to final lot testing in the production process. The utility of cell based assays which are able to investigate all three functional toxin domains, continues to be explored.

Conclusion

In conclusion, alternative methods which reduce, refine and replace the use of live animals in testing botulinum toxin therapeutic products are now available. Validation and adoption of these methods will depend on the cooperation of regulatory agencies, manufacturers and validation centres, and will ensure that the 3Rs become an accepted and routine aspect of therapeutic botulinum toxin product testing.

Acknowledgements: The authors would like to thank Dr RJ Jones and Ms Y Liu for providing some of the unpublished figures.

Glossary

ECVAM: European Centre for the Validation of Alternative Methods

FDA: Food and Drug Administration

ICCVAM: Interagency Coordination Committee on the Validation of Alternative Methods

MHRA: Medicines and Healthcare products Regulatory Agency

NIBSC: National Institute for Biological Standards and Control (NIBSC)

NICEATM: National Toxicology Program Center for the Evaluation of Alternative Toxicological Methods

Pharmacopoeia: A Pharmacopoeia is a collection of standardised specifications (monographs), which define the quality of pharmaceutical preparations and their constituents.

Parallel line method: uses a linear graph to relate the logarithm of the test doses of toxin (e.g. botulinum toxin) on the horizontal axis to the corresponding dose response scores of a biological assay (e.g. the flaccid paralysis test) on the vertical axis. The individual response scores for each dose are plotted for both the test sample and a reference standard so that the dose-response relationship for both response curves is linear and parallel and can be related to each other. Using a formula and the slope of each of the two graphs, it is possible to calculate the potency of the test sample used in the biological assay relative to the reference standard sample.

References

1. Sesardic D, Gaines Das RE & Corbel MJ (1994) Botulinum toxin: How to define biological activity. Journal of the Royal Society of Medicine (letter) 87, 307.
2. McLellan K, Gaines Das R, Ekong TAN & Sesardic D (1996) Therapeutic botulinum type A toxin: factors affecting potency. Toxicon 34, 975-985.
3. Sesardic D, Leung T & Gaines Das R (2003) Role of standards in assays of botulinum toxins: international collaborative study of three preparations of botulinum type A toxin. Biologicals 31, 265-276.
4. Hunt T & Clarke K (2008) Potency of the botulinum toxin product CNBTX-A significantly exceeds label units in standard potency test. Journal of the American Academy of Dermatology 58 (3), 517-518). 
5. European Pharmacopoeia 6.0 (2006) Botulinum Toxin Type A for injection 01/2005:2113, pp 1117-1119.
6. Sesardic D, McLellan K, Ekong TAN & Gaines Das R (1996) Refinement and validation of an alternative bioassay for potency testing of therapeutic botulinum type A toxin. Pharmacology & Toxicology 78, 283-288.
7. Sesardic D, Jones RGA, Leung T, Alsop T & Tierney R (2004) Detection of antibodies against botulinum toxins. Movement Disorders 19, 85-91.
8. Report on the ICCVAM-NICEATM/ECVAM Scientific Workshop on alternative Methods to Refine, Reduce or Replace the mouse LD50 assay for botulinum toxin testing (2008) NIH Publication No 08-6416.
9. Goschel H, Wohlfarth K, Frevert J, Dengler R, & Bigalke H (1997) Botulinum A toxin therapy: neutralizing and non-neutralizing antibodies: therapeutic consequences. Experimental Neurology 147, 96-102.
10. Rasetti-Escargueil C, Jones RGA, Liu Y, Sesardic (2009) Measurement of botulinum types A, B and E neurotoxicity using the phrenic nervehemidiaphragm: Improved precision with in-bred mice. Toxicon (in press on line).
11. Marconi S et al. (2008) A protein chip membrane-capture assay for botulinum neurotoxin activity. Toxicology and Applied Pharmacology 233 (3), 439-46.
12. Hallis B, James, BAF & Shone CC (1996) Development of novel assays for botulinum type A and B neurotoxins based on their endopeptidase activities. Journal of Clinical Microbiology 34 (8), 1934-1938.
13. Ekong TAN, Feavers I & Sesardic D (1997) Recombinant SNAP-25 is an effective substrate for Clostridium botulinum type A endopeptidase activity. Microbiology 143, 3337-3347.
14. Jones RG, Ochiai M, Liu Y, Ekong TAN & Sesardic D (2008) Development of improved SNAP-25 endopeptidase immune-assays for botulinum type A and E toxins. Journal of Immunological Methods 329, 92-101.
15. Gaines-Das RE, Heath AB, Martin H & Sesardic D (1999) Validation of in vitro assays for Botulinum Toxin: A case study. In Alternatives to Animals in the Development and Control of Biological Products for Human and Veterinary Use. (Eds Brown F, Hendriksen C and Sesardic D). Developments in Biological Standardization 101; 267-276.
16. Sesardic D (1999) Alternatives to the use of animals for bacterial toxins and antitoxins. Celebration of 50 years of progress in biological standardization and control at WHO. Developments in Biological Standardization 100, 75-82.

All views and opinions expressed in this article are those of the author and do not necessarily reflect the views and opinions of the NC3Rs.