Introduction
Implementing a rapid test method to replace currently accepted methodology for mycoplasma detection requires overcoming unique validation challenges specific to these organisms. Rapid microbiological methods (RMM) designed to detect bacteria and fungi are challenging to validate as replacements for compendial methods in general. As a result, using RMM to release finished drugs, biologics, and medical devices is relatively rare even though regulators encourage adopting these methods.
Mycoplasma testing at various stages in the production process is a regulatory requirement for certain drugs, biologics, and medical devices. The US Code of Federal Regulations 21 CFR §610.30 [1], US FDA’s Points to Consider in the Characterization of Cell LinesUsed to Produce Biologicals [2], and international compendia (USP <63> [3], EP 2.6.7 [4], JP XV General Information Chapter 14 [5]) specify accepted methodology for detection of mycoplasma. In addition to standard culture methods, the US, European, and Japanese pharmacopeias also list nucleic acid tests (NAT) as an option when properly validated; EP 2.6.7 provides a guideline with specific mycoplasma NAT validation requirements. In addition, USP <1223> [6] and EP 5.1.6 [7] give guidance on validating alternative microbiological methods; they require the alternative method to yield results equivalent or better than the compendial method. This requirement for equivalence is echoed by 21 CFR §610.9 [8] and JP XV General Notice 13 [9].
Validation Considerations
A mycoplasma test for the presence or absence of viable organisms is essentially a sterility test, suitable for many applications. Specificity and sensitivity are the most critical validation parameters for this type of qualitative test; Table 1 summarizes the additional relevant validation parameters to address for a qualitative test.
Table 1: Validation Parameters
The specificity of an alternate qualitative microbiological method is its ability to detect a range of microorganisms potentially present in the test article [6] and to unequivocally assess target nucleic acid in the presence of components expected to be present [4]. A rapid mycoplasma test should detect a broad range of mycoplasma species. It should not detect host cells or closely related bacteria.
The test method manufacturer can often provide a development summary report explaining how the assay design includes the target species, but excludes closely related species or sample matrix components. In addition, the manufacturer should provide data for a large mycoplasma panel supporting these design requirements. Assay validation at the test lab needs to establish that interference from the test sample does not produce false positive or false negative results. Negative results obtained from analysis of un-spiked test samples from a sufficient number of sample lots demonstrate interference from the test sample does not produce false positive results. Positive
results obtained from analysis of test samples spiked with mycoplasma or mycoplasma nucleic acids from representative species demonstrate interference from the test sample does not produce false negative results. The panel of species should include organisms posing the most likely contamination threats. For example, a vaccine manufacturer should probably include Mycoplasma synoviae and Mycoplasmagallisepticum if the production process uses avian material. These species may not be relevant to a therapeutic protein synthesized using CHO cells in a bioreactor.
The limit of detection is the lowest number of microorganisms in a sample detected under the stated experimental conditions [6] and the lowest amount of target nucleic acid in a sample detected but not necessarily quantitated as an exact value [4]. EP 2.6.7 indicates that replacing the culture method with an alternate method requires demonstrating a 10 CFU/mL detection limit or below. To achieve this limit, the initial sample preparation often requires a concentration or enrichment step.
The robustness of a qualitative microbiological method is its capacity to remain unaffected by small deliberate variations in method parameters [6] providing an indication of its reliability during normal usage [4]. Robustness is perhaps a validation parameter best suited to determination by the test method manufacturer because it relies upon deliberate variations in method parameters typically tested during development. The test method manufacturer can often provide a development summary report addressing various parameters, such as primer optimization or testing on many different instruments.
The precision of an analytical procedure is the agreement among individual test results for multiple samplings from a homogeneous sample. Repeatability refers to using the test procedure within a laboratory over a short time period with the same analyst and equipment [10]. Assessment requires multiple analyses using the same samples under the same conditions on the same day. Testing both un-spiked samples and samples spiked with mycoplasma nucleic acid is valuable. For a qualitative test, negative results from all un-spiked samples and positive results from all spiked samples demonstrates acceptable reproducibility. For a quantitative test, expressing reproducibility as a standard deviation or coefficient of variation may be more appropriate.
The ruggedness of a qualitative microbiological method is the precision of test results obtained by analysis using the same samples under a variety of normal test conditions, such as different analysts, instruments, reagent lots, and laboratories [6]. Assessment requires multiple analyses using the same samples under different conditions. Testing both un-spiked samples and samples spiked with mycoplasma nucleic acid is valuable. For a qualitative test, negative results from all un-spiked samples and positive results from all spiked samples demonstrates acceptable ruggedness. For a quantitative test, expressing ruggedness with equivalence statistics [11] may be more appropriate.
Finally, replacing an official method with an alternative method requires demonstrating equivalence via a comparability study. EP 2.6.7 suggests assessing all the validation parameters detailed above for both methods. One suggestion is to perform both the official method and the alternate method in parallel using the same samples. A second suggestion is to compare validation results from the alternate method to previous validation results from the official method. In addition, regulatory agencies sometimes request analysis of routine test samples using the official method and the alternate method run in parallel as a condition of approval. Providing this data with the validation package may address this request proactively.
Number 10: Method Selection
There are several things to consider prior to selecting a rapid mycoplasma test. It is critical to first define the requirements for the proposed application. Tests for product release, in-process monitoring, raw materials, and cell line qualification have different regulatory implications with different validation requirements. EP 2.6.7 suggests method validation alone is adequate when using the test as a complementary test (in-process monitoring); a comparability study is necessary when replacing an official method with an alternate method (product release). A risk-based approach consistent with ICH Guideline Q9 is valuable when selecting the appropriate technique to use for a specific application [12]. In the past year, commercial manufacturers introduced rapid mycoplasma tests claiming to meet the validation requirements in EP 2.6.7. The tests all detect mycoplasma nucleic acids, but use different techniques to achieve optimal specificity and sensitivity. These techniques include real-time PCR, touchdown PCR, transcription mediated amplification, microarrays, and hybrids using growth and PCR. In addition to providing adequate specificity and sensitivity, some rapid tests also offer quantitation and identification. These attributes may be important considerations, especially for conducting timely failure investigations. They may not be necessary however, if the contamination rate is exceedingly small or if they add significant cost to each test. Timing is also a concern. In some situations, such as autologous cell therapy product release or inprocess contamination control, it is imperative to obtain test results as quickly as possible, often the same day. For other applications, especially those currently requiring a 14-day sterility test for bacteria and fungi, a longer test may be adequate.
Number 9: Strain Characterization
Mycoplasmas are a complex and unique group of bacteria belonging to the class Mollicutes are characterized by their permanent lack of a cell wall, a small genome size (0.58 – 2.2 Mbp) and a low G + C content (23 – 40 mol%) in their genomes [13]. These flexible, pleomorphic organisms can be as small as 0.2 – 0.3 μm and can achieve very high densities in cell cultures (107 – 108 organisms/mL) without discernable changes in pH or turbidity [14]. There are currently more than 200 known species in 9 genera containing many pathogens. The majority of cell culture contaminants belong to only 6 species of human, bovine, or porcine origin: M. hyorhinis,M. arginini, M. salivarium, M. orale, M. fermentans, and A. laidlawii. In contrast to testing for bacteria and fungi, testing for mycoplasma is particularly challenging because they are not cultured and detected using traditional microbiological techniques. Preparing samples containing a well-characterized quantity of correctly identified viable organisms requires specialized techniques. Contract laboratories specializing in mycoplasma culture, identification, and testing can provide expertise in this area.
Number 8: Nucleic Acid Detection
Nucleic acid detection is not necessarily equivalent to organism detection. Culture methods detect only viable organisms, while commercially available rapid mycoplasma tests primarily detect mycoplasma nucleic acids. The nucleic acids may come from viable organisms, non-viable organisms, or lysed organisms. In addition, a single organism may contain multiple nucleic acid copies. Both assay design parameters and validation experiments should characterize the discrepancy between organism detection and nucleic acid detection.
Number 7: Non-viable Organism Detection
Tests designed to detect mycoplasma nucleic acids may detect non-viable organisms in addition to viable organisms. Detecting mycoplasma nucleic acids in a sample containing only non-viable organisms could generate false positive results leading to unnecessary failure investigations or product rejections. Methods based on RNA detection may address this issue by detecting primarily organisms actively undergoing protein synthesis. Quantitative methods based on DNA detection may address this issue by quantifying mycoplasma DNA concentration over time, with increasing concentration indicating replicating organisms.
Number 6: Mycoplasma “CFU” Interpretation
Mycoplasma “colony forming units” can be difficult to interpret because one “CFU” may contain multiple organisms. Aggregation, filamentous morphology, inadequate growth media, and growth differences on the host-cell surface, in broth, and on agar makes differentiating a single viable organism as a colony forming unit challenging [15]. Mycoplasma culture requires a specialized skillset above and beyond basic microbiology training. Consulting mycoplasma testing experts is advisable to minimize errors introduced from improper handling and incorrect colony assessments.
Number 5: Nucleic Acid/CFU Ratio
The nucleic acid/CFU ratio expresses the DNA or RNA copy number contained in each species. Determining the nucleic acid/CFU ratio is important to establish whether the alternative method yields results equivalent or better than the compendial method. It is important to reduce the nucleic acid contribution from non-viable organisms. Using stock solutions containing high concentrations of non-viable organisms biases detection in favor of nucleic acidbased methods. Selecting fastidious strains, minimizing carryover from the initial inoculum, and using actively growing cultures helps reduce this contribution [15]. Establishing the detection limit in terms of nucleic acid copies requires spiking the test system with 10 mycoplasma nucleic acid copies or below. Purified nucleic acids from representative species facilitate these assessments. It is possible to establish the detection limit solely using nucleic acids. In this case, establishing the relationship between colony forming units and nucleic acid copies is necessary, requiring a quantitative technique.
Number 4: Small Sample Volume
Most detection methods based on nucleic acid amplification introduce small sample volumes (less than 100 μL) during the amplification step. Since the culture methods often require as much as 15 mL, this discrepancy could result in as much as a 100-fold loss in sensitivity due to sample volume alone when comparing the methods. Adding a concentration step or an enrichment step to the sample preparation may overcome this loss in sensitivity. After appropriate sample preparation, many alternative methods may approach the theoretical 1 CFU/mL culture method detection limit. Establishing the detection limit requires inoculating the test sample with 10 CFU/mL or below using challenge organisms. Mycoplasma stock suspensions from representative species, previously enumerated via plate count or quantitative methods, can facilitate inoculum preparation.
Number 3: Low Inocula
Technical limitations become a factor when inoculating low microorganism concentrations (<10 CFU). Preparing a bacterial inoculum with a uniform cell count per unit volume has significant error associated with sampling, dilution, plating, incubation, counting, and calculation [16]. Difficulties distinguishing a single mycoplasma organism as a colony forming unit can compound these errors. In contrast, preparing an accurate standard solution containing a known nucleic acid copy number by dilution is relatively straightforward. Analyzing a large number of samples is one approach to address statistical variability at low inocula. Quantitatively determining the nucleic acid/CFU ratio and enumerating the spike solutions is another approach.
Number 2: Gold Standard Method Validation
Rigorous assay validation data for the direct culture and indirect DNA fluorochrome staining methods is often not readily available. This is important when designing comparative studies to show equivalence. Theoretically, the combination of these methods detects the broad range of species at concentrations as low as 1 CFU/mL. As a result, comparative studies may have to demonstrate equivalence to these theoretical criteria. These theoretical criteria may be more stringent than those established through validation.
Number 1: Implementation in a Cell Culture Production Facility
Introducing viable mycoplasma cultures into a cell culture facility producing therapeutic products may present an unacceptable contamination risk. Contamination routes include materials, personnel, donor tissue, and cross-contamination. Common disinfectants, such as isopropyl alcohol, effectively lyse mycoplasma. Contamination controls employed in a cGMP production facility should reduce the risk of a widespread contamination event. Even so, most cell culture facilities choose to eliminate any risk from test lab personnel contaminating production cultures by banning viable mycoplasma cultures in the entire facility, including the test lab. This presents significant logistical challenges for validation, technology transfer, and ongoing training. Validating the assay off-site using a contract facility is the least difficult challenge to overcome. In fact, it is probably desirable based on the specialized skillset needed to work with mycoplasma. Subsequent technology transfer and analyst qualification at an off-site contract facility is more difficult due to travel, expense, and logistics. It also may be unnecessary because nucleic acid spikes should demonstrate acceptable technology transfer and training. Demonstrating technology transfer and training for initial concentration steps where the physics is well understood, such as for centrifugation or membrane filtration, seems superfluous. If required, developing surrogates for mycoplasma detectable by the assay could address the issue.
Conclusion
Validation considerations specific to rapid mycoplasma testing can impact successful implementation. These include
10. Selecting an appropriate method
9. Characterizing mycoplasma strains
8. Discriminating between nucleic acid detection and organism detection
7. Minimizing non-viable organism detection
6. Interpreting mycoplasma “colony forming units”
5. Determining the nucleic acid/CFU ratio
4. Addressing small nucleic acid amplification sample volumes
3. Overcoming errors associated with preparing low inocula
2. Obtaining validation data for official methods
1. Reducing contamination risk in a cell culture facility
Proactively addressing these issues prior to validation should facilitate implementing rapid mycoplasma test methods at appropriate stages in the production process for certain drugs, biologics, and medical devices.
References
1. Test for Mycoplasma. In: Code of Federal Regulations. Title 21 §610.30; 2009.
2. Recommended Procedures for Detection of Mycoplasma Contamination in Biological Products Produced in Cell Substrates. Points to Consider in the Characterization of Cell Lines Used to Produce Biologicals. Rockville, MD: The United States Food and Drug Administration Center for Biologics Evaluation and Research; 2003.
3. <63> Mycoplasma Tests. In: United States Pharmacopeia. 33rd ed. Rockville, MD: The United States Pharmacopeial Convention, Inc.; 2010.
4. 2.6.7. Mycoplasmas. In: European Pharmacopoeia. 6th ed. Strasbourg, FR: European Directorate for the Quality of Medicines; 2010.
5. Mycoplasma Testing for Cell Substrates used for the Production of Biotechnological/Biological Products. In: Japanese Pharmacopoeia. 15th ed. Tokyo, JP: Ministry of Health, Labour and Welfare; 2006.
6. <1223> Validation of Alternative Microbiological Methods. In: United States Pharmacopeia. 32nd ed. Rockville, MD: The United States Pharmacopeial Convention, Inc.; 2009.
7. 5.1.6. Alternative Methods for Control of Microbiological Quality. In: European Pharmacopoeia. 6th ed. Strasbourg, FR: European Directorate for the Quality of Medicines; 2010.
8. Equivalent methods and processes. In: Code of Federal Regulations. Title 21 §610.9; 2009.
9. General Notices. In: Japanese Pharmacopoeia. 15th ed. Tokyo, JP: Ministry of Health, Labour and Welfare; 2006.
10. <1225> Validation of Compendial Procedures. In: United States Pharmacopeia. 32nd ed. Rockville, MD: The United States Pharmacopeial Convention, Inc.; 2009.
11. Chambers D, Kelly G, Limentani G, et al. Analytical Method Equivalency: An Acceptable Analytical Practice. Pharm. Technol. 2005; 29:64-80.
12. Duguid J, Kielpinski G, Seymour B, du Moulin GC. Risk Assessment for a Rapid Mycoplasma Test Optimized for Cell Therapy Products. Am. Pharm. Rev. 2009; 12:100-104.
13. Bacterial Pathogens. Mycosafe Diagnostics GmbH Web site. 2008. Available at: http://www.mycosafe.at/mycosafe/mycosafe.nsf/alldocs/7C21DDE60B3859B3C125700E00576B02?OpenDocument. Accessed August 5, 2008.
14. Duguid J, Kielpinski G, du Moulin GC, Seymour B. Application of a Risk- Based Approach to Optimize a Rapid Mycoplasma Test for Cell Therapy and Tissue-Engineered Products. Atlanta, GA: AAPS Annual Meeting and Exposition; 2008.
15. Windsor H, Windsor D. NAT vs Microbial Culture: Comparing Chalk with Cheese? Gaithersburg, MD: FDA/CBER - Public Workshop: Rapid Methods for Detecting Mycoplasma Contamination in the Manufacture of Vaccines, Including Pandemic Influenza Vaccines, and Other Biological Products; 2008.
16. Technical Report No. 33: Evaluation, Validation and Implementation of New Microbiological Testing Methods. PDA J. Pharm. Sci. Technol. 2000; 54:23.
John Duguid is a scientist at Genzyme responsible for developing and implementing rapid microbiological assays. Previously, he managed QC cell therapy operations. Mr. Duguid received his BS in Chemistry from the University of Michigan. Prior to Genzyme, he worked as an analytical chemist at Abbott Laboratories and Arthur D. Little. Readers may contact the author directly at: [email protected]