Risk assessment and risk management of transgenic pharmaplants

I wanted to take the opportunity to discuss what seems to be important consideration in the risk assessment of some—perhaps in time the majority—of pharma plants: The use of chloroplast genetic engineering. The use of transplastomic plants for drug production represents a departure from more “traditional” nuclear genome engineering and may raise some unique considerations for risk assessment. While some of these features are not unique to pharma plants per se, the nature of the target and marker transgenic proteins as bioactive macromolecules deserves particular attention.

Some relevant examples:


HGT potential

Chloroplasts are transformed via homologous recombination, typically via particle bombardment, and utilize bacterial-like border sequences to facilitate site-specific integration of heterologous cassettes (given that chloroplasts genomes former Cyanobacteria). Whether this would alter the transformation frequency of horizontal gene transfer to these microbes in the environment may be important in some cases.


High(er) concentrations of heterologous proteins

Transplastomic plants have been reported to achieve protein expression rates of up to 45% total soluble protein (TSP) of the target protein and 10% TSP of marker genes. The implication of the potential metabolic cost, and hence fitness to the recipient host, along with the impact of greater concentrations of bioactive proteins in the environment may garner consideration in specific uses. The prudence of limiting marker genes in the environment, particularly with antibiotic resistance genes, where possible has been roundly advocated and should follow with strategies to minimize their use in transplastomic plants.


Consistency of maternal inheritance

One widely touted benefit of transplastomic plants is may limit pollen transfer of transgenic sequences, as chloroplast genomes are only maternally inherited  (though we all know too well in biology there are few universals!). However, the “leakiness” of this system has been experimentally show to be often variable, and often species dependent. Therefore an assessment of this “leakiness” potential may be warranted in some cases.


RNA editing, post-translational modification, and affinity tag
related changes in primary and secondary structures of transgenic protein expression

There is evidence that the complex RNA editing of chloroplast transcripts are critical for faithful expression, and translation of unedited exogenous messenger RNAs in transplastomic cassettes could lead to functionally defective proteins. Beyond these transcriptional differences, post-translational modifications are often quite different depending on host. Plants particularly are known to add that may alter it allergenicity (addition of fucose and xylose N-linked glycans) or desired functionality or stability (due to folding or structural differences), all would need to be clarified through validated methods. Lastly the use of affinity tags, such as his or FLAG epitope tags used to facilitate efficient protein recovery and purification from the plant require changes in protein primary structure.  A molecular characterization of the insertion site, along with primary and secondary structural features of the target protein and related to potential functional changes, may be necessary for some classes of protein or production systems.


Co-transfer of heterologous integration cassettes to the nucleus

Certain transformation methods, notably Agrobacterium-mediated gene transfer and particle bombardment (the two most common means of transformation) putatively co-transfer integration cassettes into the nucleus in some cases. This represents a distinct regulatory challenge that can be best addressed through localization studies in each transformation event under assessment.


Transgene stability

The maintenance of homoplasmy among the thousands of copies of chloroplast genomes represents one of the big challenges for stable expression of chloroplast-transformed transgenes. Transgene stability may need to be demonstrated, in specific cases, for regulatory approval.

The above scenarios represent some of the known uncertainties surrounding the use of transplastomic plants for the expression of heterologous DNA. As this method is becoming widely used in the production of pharma plants, well formed questions and validated research methodologies can strengthen the scientific quality of assessments where chloroplast engineering may bring new risk scenarios.

-David Quist

I would like to add to Dr. Quist's comments with regard to chloroplast engineering, with the associated potential use for containment, and Dr. Tappeser’s comments about the need for effective containment of transgenes. All references may be found here: ftp://ftp.fao.org/ag/cgrfa/bsp/bsp35r1e.pdf

Neither the US National Academy of Science nor a recent UN FAO report on transgene flow found it realistic, even with the use of hypothetical sterilization technologies, to expect that technology will be sufficient to stop transgene flow. Possibly no combination of technology and human management may be sufficient either. Therefore, for the foreseeable future, only contained laboratories can serve to manage the risks from GMOs that may cause unacceptable harm if the transgene were to escape.

I’d like to offer some comments on both 794 and 834.

In re 794, Dr. Quist makes some interesting points, but fails to acknowledge some overarching facts of salient importance.  There are no uncertainties or safety issues that the current processes of risk assessment and management are incapable of dealing with; indeed, a number of them have been dealt with in the context of transgenic plants not intended for use in pharmaceutical production from the very beginning, e.g., the issues of stability in construct inheritance, potential fitness impacts of the transgenic modification, potential for horizontal transfer, species specific impacts of various sorts, and so on.  Nothing here is novel or unfamiliar.

A more fundamental issue not acknowledged either in posts 794 or 834 is that of hazard identification – both postings seem to assume a necessarily higher level of intrinsic hazard from PMP plants than from other transgenic (or conventional) plants.  This is not a reliable assumption, as long experience has shown with conventional plants from which pharmaceuticals have been derived (vincristine and vinblastine from Madagascar periwinkle; salicylic acid from Salix spp., Vitamin E from soya or C from citrus, etc.) and on to components from mother’s milk in modern transgenic plants, such as lactoferrin, in rice.  It would not be a good and reasonable investment of resources, nor a scientifically defensible approach to risk assessment and management, to treat all PMP plants as if they are potentially highly hazardous, and regulators and policy makers should not lose sight of such facts.

Even with plants making pharmaceuticals that are demonstrably safe, however, market disruptions can occur if they wind up where not desired.  Food companies are concerned about their brands which may suffer if unapproved materials (safe or not) are found in them, rendering them “adulterated” under law.  For this reason many PMP companies are trying to work with non food plants, but even those using food plants are devoting unprecedented attention to identity preservation, with considerable isolation distances, dedicated machinery, and other methods of containment.  No company would rely solely on biological containment, as many biological measures can be, under certain circumstances, less than absolute, as noted in 834.  This is why developers of PMP routinely rely on a defense-in-depth approach using multiple different methods intended to contain PMP materials and keep them out of places they are not wanted.

It is also the case that for the foreseeable future regulatory agencies will continue to treat each and every PMP field trial request with focused, case-by-case attention to ensure materials are properly handled without surprises.

But in all of these cases, the potential for PMP caused impacts on biodiversity are effectively indistinguishable from zero.  While agriculture itself has huge impacts on biodiversity (primarily through the loss of native lands converted to agricultural production) the areas involved in PMP production will be, at best, miniscule by comparison to the land devoted to agriculture in general.  The extraordinary added value of PMP plants will ensure high levels of containment and special handling likely to mitigate or eliminate any significant hazard to biodiversity.  On any list of the top threats to biodiversity, no serious scholar of biodiversity would place PMP plants.  There are many more real and present dangers.