ISSUES RELATING TO ON-SITE SAMPLING AND ANALYSIS

Dr. Jack Melling

Director, The Salk Institute Biologicals Development Center

August 19, 1997

1. Sampling and Analysis for Detection versus Identification

Whereas full identification of a sample taken during on-site activities under a BWC compliance regime might raise serious confidentiality problems, there are clear advantages to on-site sampling and analysis activities which are targeted solely at confirming the absence, or presence, of certain agents. Such blinded and targeted procedures are available, and would eliminate the risk of loss of proprietary information from that part of the inspection procedure, while giving reassurance that a facility was not working on certain specified agents.

2. Sampling

Technology exists to take samples at almost all stages of pharmaceutical processing and to carry out analysis in real time. The possible exception would at the final freeze-drying or container-filling stages, where there is special and real concern about potential contamination of the product and the plant facilities that handle it. At other stages there are sampling points normally used for quality control, at which samples can be taken for analysis without detriment to the manufacturing process.

3. Assay Technology

Two principal types of analysis are available for either Yes/No detection or more detailed characterization: immunological assays and DNA assays. New analytical methods presently under development will eventually provide additional useful procedures for blinded detection of microorganisms.

There are two components involved in sample analysis: the "hardware", i.e., the apparatus, and the "software", i.e., the specific chemical reagents. The specific apparatus to be used will depend on which type of assay is chosen and on the speed and certainty required. For both types of assay, there are equipment formats commercially available which, subject to appropriate validation, are likely to be acceptable. These are portable and the simplest may not even need a power supply.

For simple Yes/No detection of an agent, immunological analysis can be employed using monoclonal antibodies, which react with only a single antigen of the many possessed by a pathogen -- i.e., they recognize only one of its unique biochemical components. A variety of monoclonal antibodies has been produced for recognizing a variety of antigens specific to a number of pathogenic organisms. Although at present there may not be a complete range of monoclonal antibodies available to cover all the agents of concern, there should be no insuperable problem in developing the necessary array, especially given current efforts to develop techniques for BW detection.

It is important that an inspection team should not be able to "fingerprint" vaccines or antisera to be used for prophylaxis or treatment against BW agents, because that information could be used to tailor agents so as to evade protection or detection systems.

This problem can be eliminated be making sure that any immunological reagent to detect a particular organism contains several different monoclonal antibodies, each of which will recognize a different biochemical constituent, or antigen, of the organism. In this way, a positive result would not divulge which biochemical substance were present in the sample, or the amounts of those substances. For example, let's assume we have an organism which has ten biochemical components that could identify it, and that one of those components can be isolated for use as a vaccine. A panel of ten monoclonal antibodies, each able to recognize one of the ten biochemical components, will give positive results for samples taken at all stages during the isolation and purification of the vaccine component, but there will be no way of knowing which of the ten is in the final vaccine.

Similarly, in the case of DNA analysis, the short DNA sequences used as primers for amplifying and probes for identifying genetic sequences in a sample can be selected to target genes that are characteristic of the species and to avoid those that would identify strains or variants, where intellectual property issues would most likely arise.

The DNA primers and probes and the monoclonal antibodies, as well as the equipment, would have to be standardized and validated in advance for use in sample analysis under a BWC compliance regime.

Using these methods, inspectors at a facility claiming to employ a human gene cloned in the bacterium E. coli to produce insulin, for example, could verify that the organisms contained no toxin genes/produced no toxin, and that no anthrax or other microbial pathogens were present, but they could not learn anything about the insulin gene or other genetic components of the E coli.

Non-virulent strains of potential BW agents could be verified by several means, including direct testing of the agent on animals to see if it caused disease -- thereby protecting all other information about the strains.

4. Preventing the Removal of Organisms

A real concern, in international inspections, is the opportunity for theft of valuable strains. This could be deterred, as it is on a national basis, by the adoption of penal sanctions. Carrying out analysis on-site would be an important safeguard. In addition, requiring inspectors to undergo a complete change of clothing and "shower-out" of facilities would do much to reduce concerns. Another option would involve treatment of samples by facility personnel to kill microorganisms, and/or partially degrade their DNA (with a restriction enzyme), before analysis by inspectors.

Decontamination of any equipment which may have come into contact with microbial or biochemical samples would also be necessary, but could be avoided by the use of disposable items wherever possible. Appropriate equipment has been developed.

5. Conclusion

It is possible to establish a sampling and analysis regime that would reliably safeguard confidential information while providing compliance information. Because sampling and analysis procedures are technical and can be precisely defined, they are more controllable than any other on-site measure. Thus, sampling and analysis could, in some circumstances, qualify as the least intrusive means for demonstrating compliance.