Herbal Authenticity as per DNA Barcoding in Canada
Abstract & Commentary
By David Kiefer, MD, Editor
Synopsis: A DNA identification technique led this research group to postulate that many herbal products and herbal leaf samples were incorrectly identified and/or adulterated.
Source:Newmaster SG, et al. DNA barcoding detects contamination and substitution in North American herbal products. BMC Med 2013;11:222.
- Approximately half (48%) of 44 herbal products tested had the correct plant DNA present, but more than half of the products (59%) contained at least some plants not listed on the label.
- Fillers and plant adulterants were also found.
- There are some concerns about the methodology and clinical relevance of these results.
The researchers of this study used a biotechnology technique, DNA barcoding, to analyze natural product samples, looking for short, standardized gene sequences specific to a given species. Multiple DNA extracts were taken from 100 mg of the herbal products and analyzed using standard polymerase chain reaction techniques for two gene sequences, as per references mentioned in this article. The two sequences used were rbcL (imperfect for separating the identity of closely related species) and ITS2 (high species resolution based on nuclear, rather than plastid, genome). The reference for these samples was called a standard reference material herbal barcode library (currently at 100 species, but being expanded by the researchers), which contains known gene sequences to which the samples can be compared. Of note, as per a critique that appeared soon after the publication of this article, this study's reference genetic library was not the same as that used as an herbal industry standard.1
The researchers blindly chose 44 single-ingredient herbal products from 12 companies. These were 41 capsules, one tablet, and two powders, representing 30 different species of plants. These were submitted for analysis using only a product label; brand names were excluded. The herbal products chosen were brands commonly available in North American retail outlets and were purchased from stores in Toronto, as well as through mail order from the United States.
Additionally, 50 leaf samples, representing 42 plant species, were collected from taxonomically identified (the "gold standard") plants in horticulture greenhouses, accounting for all 30 of the herbal product plant species as well as related species. These were also blindly submitted for analysis as an independent test of validity of the DNA barcoding technique.
The samples were then categorized as authentic (DNA matched the species listed on the label), contamination (DNA found for a species other than what was listed on the label), substitution (DNA found for a species other than what was listed on the label, and the label species DNA was not found), and filler (DNA from known fillers, such as wheat, soybean, and rice, was detected).
The results were as follows. First, the DNA from the 100 reference plants showed 100% species resolution, in that there were DNA regions specific to each of those plants as analyzed by this technique. Of the herbal products, 40/44 (91%) were successfully DNA barcoded using the two combined gene regions. The more accurate gene region, ITS2, only missed three herbal products, which the researchers posit was due to substitution by plants not in their reference library. For the herbal leaf samples, 57% were matched correctly to the reference library using rbcL and 100% using ITS2.
The categorical breakdown of the herbal products is shown in Table 1. The researchers combined these results by pointing out that 59% of the products contained plant species not on the labels, while 33% also contained plant fillers or contaminants not listed on the labels. Approximately 8% of the products contained unidentifiable sequences not yet in the reference library; some of these were sequences represented plants in completely different plant families from the plants on the labels.
The researchers provided company-by-company data as well. Only two of the 12 companies had only authentic products (10/12 with product substitution), whereas three companies had no authentic products. However, for 90% of the plant species analyzed, the researchers found at least one company that "got it right," that is, had an authentic product.
A full table was provided for the product labels as compared to the DNA barcoding findings, with some interesting, if not concerning, results. For example, one Ginkgo biloba product was actually black walnut (Juglans nigra), while only rice was found in one Ginkgo biloba product and one St. John's wort (Hypericum perforatum) product.
The headlines were striking: Almost every major newspaper had some version of the New York Times warning "Herbal Supplements Are Often Not
What They Seem."2 Dear to the heart of dietary supplement users, these results fanned the fire of anxiety and fear about supplement quality. Additionally, if this author (DK) is a litmus test of reality, this study may have been the first time that many people heard of the use of DNA barcoding, presumed the "gold standard," for everything from forensics to "personalized medicine," and now, apparently, for herbal product identification. The story, however, is not so simple and clear-cut, primarily because of some methodological problems and a technology that is still in its infancy.
On this note, the paper brought the ire of herbal industry experts, pointing out numerous inconsistencies and flaws. Most notably, and probably a fatal oversight, is that DNA may not be present in herbal extracts, even though the physiologically active phytochemicals will be. Without a product list, it is difficult to know how many of the 44 products tested were standardized extracts and therefore not expected to "light up" on a DNA barcoding test. Along the same vein, the heat treatment that most plants undergo in processing for consumer sale will damage, if not destroy, DNA, hopefully leaving the phytochemicals intact. Again, no DNA would expected to be present. Finally, the polymerase chain reaction technique, part of this DNA barcoding analysis, is very sensitive for even the faintest trace amounts of DNA; contamination in the testing laboratory, not the herbal products, can't be ruled out as a cause of the "contamination" category of these findings. Proper controls, lacking in this study, could have theoretically mitigated these false negatives.
The authors of this study mention some of the prior work on DNA barcoding, and marketplace analyses of adulteration and substitution, which are in the ballpark for the percentages presented here. They aptly put their results into clinical context, with some of the dangerous substitutions at the top of their discussion, including the presence of wheat as a filler, possibly of concern to our patients with celiac disease. Fillers by definition aren't "bad," and actually are necessary (and legally approved) to help formulate products. Consumers have the right to know what's in each product, and such compounds should be on the label. Alternatively, they mention some of the pitfalls of their work including the incomplete herbal DNA reference library.
Strict attention to herbal quality is of paramount importance, but it remains to be seen whether this DNA barcoding technique is the best way to assess authenticity. It needs to be corroborated and cross-checked with other laboratories and gene banks. The study mentions several papers and sources for other herbal reference libraries.1 Insomuch as the researchers themselves state that "There are currently no best practices in place for identifying the species of the various ingredients used in herbal products," a combination of adherence to the legally mandated current good manufacturing processes and refinement of the DNA barcoding techniques might be just the perfect combination to foster product quality.
- Gafner S, et al. ABC Review and Critique of the Research Article "DNA Barcoding Detects Contamination and Substitution in North American Herbal Products" by Newmaster et al. HerbalEGram November 2013. Available at: http://cms.herbalgram.org/heg/volume10/11November/DNAbarcodingReviewandCritique.html?t=1383684796. Accessed November 30, 2013.
- O'Connor A. Herbal Supplements Are Often Not What They Seem. NY Times November 3, 2013. Available at: http://www.nytimes.com/2013/11/05/science/herbal-supplements-are-often-not-what-they-seem.html?_r=0. Accessed November 30, 2013.