A practical guide for transparent reporting of research on natural products in the British Journal of Pharmacology: Reproducibility of natural product research

Citations Over TimeTop 1% of 2020 papers

Abstract

Natural products continue to be an important source of medicines and drug templates. Indeed, it has been estimated that approximately a third of all Food and Drug Administration (FDA)-approved drugs over the past 20 years are based on natural products or their derivatives (Thomford et al., 2018). This is likely to be due to the vast chemical diversity of natural products, which increases the probability of finding structurally distinct ‘lead compounds’ for different targets and diseases (Gu et al., 2013). For instance, the Dictionary of Natural Products has thus far recorded ~200,000 plant secondary metabolites, including about 170,000 unique structures (Harvey, Edrada-Ebel, & Quinn, 2015). The diversity and growing opportunity is also reflected in the size of natural product libraries that have been created, such as Supernatural II (http://bioinf-applied.charite.de/supernatural_new/index.php?site=home), which has 325,508 different natural compounds (Banerjee et al., 2015), and the open access initiative (led by the US National Cancer Institute) with libraries of ~80,000 plants, ~20,000 marine samples and ~6,000 microbes. This latter initiative estimates that ~1,000,000 distinct fractions will be derived over the next 3–4 years (https://dtp.cancer.gov/organization/npb/npnpd_prefractionated_library.htm). In addition to its well-established role in drug discovery, natural product pharmacology is relevant in the context of dietary supplements, which are part of the vast and nebulous nutraceutical market and includes vitamins, amino acids, proteins, minerals, fibres, plant extracts and natural compounds (Andrew & Izzo, 2017). Due to a lack of rigorous regulation in many countries, dietary supplements can be marketed without clinical evidence of efficacy and with uncertain composition (Andrew & Izzo, 2017). Despite these shortcomings, demand for dietary supplements continues to grow (Williamson, Liu, & Izzo, 2020). Herbal dietary supplement sales in the United States experienced record growth in 2018 and consumers spent $8.842 billion on herbal supplements across all market channels (Smith, Gillespie, Eckl, Knepper, & Reynolds, 2018). Furthermore, traditional medicine (such as Ayurvedic medicine, traditional Chinese medicine and Unani) not only continues to form an integral part of treatment within certain cultures but has also entered Western culture (Williamson et al., 2020). With this background, it is not surprising that BJP receives many submissions that focus on the pharmacology of specific natural products. For the great majority of natural product papers published in BJP, the authors include final summative statements that propose a naturally occurring molecule for further clinical development or immediate use as a dietary supplement in the nutraceutical market. Many of these papers also identify novel drug targets, elucidate the mode of action, discover lead compounds and offer potential repositioning of natural products already in clinical use (see Table 1 for a list of examples). All of these issues are highly relevant to the BJP and offer novel pharmacological approaches to discovery science as well as potential therapies. BJP does not publish papers that are limited to identification of natural products. Isolation, purification, elucidation of structure and semi-synthesis of chemical compounds are considered only in a context in which both a robust, deep and detailed pharmacological analysis and mechanism(s) for biological activity attributed to the natural product are provided. As an example, in a BJP paper published by Yin et al. (2015), the authors provided evidence of the synthesis of a previously isolated metabolite from Danshen (Salvia miltiorrhiza), which was identified using MS and proton and carbon NMR spectra. In addition, the authors provided a detailed pharmacological analysis, which showed that the natural compound prevented isoprenaline-induced cardiac fibrosis by inhibiting a NOX2/ROS/p38 pathway. Identification of the possible mechanism of action of the natural product is mandatory for publication in BJP. BJP considers only papers that describe studies on purified active compounds; in such cases, the purity of the compound, as well as major impurities, must be reported. BJP does not consider papers on mixtures of compounds (such as herbal extracts) alone, unless these observations are accompanied by evidence identifying and demonstrating similar activity of a purified component(s) of that extract. In some cases, the activity of an extract or mixture may be due to additive or synergistic activities of several compounds. If this is the case, then this finding should be shown by combining purified constituents and demonstrating functional activity that replicates that of the mixture. Overall, it is important to define which, and how, compounds of the extract contribute to the pharmacological activity. As an example in which evaluation of an extract was followed by a detailed pharmacological analysis of the main active ingredient, a BJP paper reported that multiple distinct Chinese herbal medicine extracts increased the expression of uncoupling protein 1 in isolated adipocytes (Nie et al., 2018). The authors found that extracts from the Chinese plant Astragalus membranaceous had the highest activity and then went on to isolate and purify the components of that extract (with data from HPLC identifying the constituents). The authors identified the isoflavone formononetin as the chemical component responsible for the functional activity. Subsequently, the authors elucidated the pharmacological profile and the mode of action of formononetin, using a number of pharmacological and molecular approaches, clearly showing its binding to PPARγ and that it blunted weight gain and increased energy expenditure in obese mice. Full details regarding the provenance and processing of extracts must be provided. These will include the part of the plant/marine entity or microbe used, the method of extraction, the yield of dried extract as a percentage weight of the starting fresh or dried material, and type and concentration of extraction solvent. For all starting materials, the formal biological name should be given from which the extract is derived. The scientific name (including the family) should be used, a according to the Integrated Taxonomic Information System website, and should be in the binomial format composed of the genus and species. There are instances where naming is not straightforward, in which case full explanations and justifications should be provided. The source of the product (i.e., country and region, from the wild or cultivated) should be stated. For rare organisms and all plants, a logged sample/voucher specimen stored within an accessible database in a recognized institution must be provided to enable others to access the sample and conduct analyses (Culley, 2013). Samples held in personal and private stores that are not available to other scientists are not considered acceptable for publication in BJP. To ensure reproducible pharmacological activity, the extract must be chemically characterized (e.g., by HPLC fingerprint and metabolomics) and the content of marker compound(s) must be measured with validated analytical methods. For extracted compounds, phytochemical characterization, purity (%) and methods used to determine compound identity and purity must be stated. Most compounds should be tested at high purity (95–99%). In some instances, this may be difficult to achieve and, if so, authors are required to provide a clear description of the other constituents and explain how these other components have been accounted for in the analyses. A figure that shows the structure of the extracted compound must be included in the manuscript. For commercially available compounds, the name of the supplier must be given. Overall, reproducibility of findings is facilitated by the full disclosure of source of plant/tissue (provenance) and extraction process. To enable the latter, BJP has no word count restriction on methods. All extracts used must be available to readers either via an open access facility or from the authors. Many compounds of natural origin are not easily soluble in water or saline, and hence, they may be dissolved in organic solvents, such as DMSO, ethanol, vegetable oils (sesame oil and corn oil) or ethanol. Unfortunately, many organic solvents, at relatively low concentrations, affect cell lines or isolated tissues in ways that affect outcomes of in vivo investigations. Cell-based assays are often intolerant to solvent concentrations of greater than 1% (Hughes, Rees, Kalindjian, & Philpott, 2011). For example, DMSO, which is probably the most frequently used organic vehicle, exerts a number of pharmacological actions including differentiation of malignant cells, antioxidant and antibacterial activity, vasodilatation and smooth muscle relaxation, and behavioural and in vivo cardiovascular effects (Castro, Hogan, Benson, Shehata, & Landauer, 1995; Jacob & Herschler, 1986; Parisi et al., 2010). Therefore, we advise that assessment of the minimum concentration of vehicle required to enable solubilization is conducted and mandate that the effect of the vehicle on the responses under study is reported. Inadequate consideration of the effects of the vehicle can lead to incorrect judgement regarding responses to the intervention. A recent example of incomplete consideration of vehicle effects occurred in studies of the cardiovascular actions of marine lipids. A substantial literature has reported positive effects of marine lipids on circulating triglyceride levels. Raised plasma triglycerides correlate with worse outcome in terms of cardiovascular events (Sarwar et al., 2007). Results for a large prospective study assessing the effects of icosapent ethyl in patients with increased triglyceride levels (REDUCE-IT) was recently published (Bhatt et al., 2019). The investigators suggested that icosapent ethyl reduced the cumulative incidence of cardiovascular events compared to the placebo (a mineral oil mimicking the colour and consistency of the fish oil), However, LDL cholesterol (LDL-C), while similar in both groups at baseline (750 mg·l−1), increased by 20 mg·l−1 in the active intervention group but by 70 mg·l−1 in the placebo (mineral oil) group. This level of LDL-C is known to be associated with an increase of events and thus raised the possibility that the icosapent ethyl might be simply preventing the detrimental effects of another component of the fish oil. There are also instances when, in order to reach high concentrations of the compound under investigation (e.g., to construct a full concentration–response curve in agonist/antagonist studies), the vehicle can affect the response being studied. In such cases, concentration–response curves for both the vehicle and the compound under investigation must be compared and shown. As an example, in BJP, Thomas et al. (2007) reported that the plant-derived compound cannabidiol displayed high potency as an antagonist of CB1 and CB2 receptors. The [35S]GTPγS binding assay was used to determine both the efficacy of cannabidiol and the ability of cannabidiol to antagonize cannabinoid receptor agonists at the mouse CB1 and the human CB2 receptor. In this assay, the concentration–response curve related to the vehicle (DMSO) was clearly shown in figures and enabled comparison with the curve related to the effect of cannabidiol in the vehicle. Commonly used pharmacological screening assays for new compounds, isolated from natural sources, are cell culture systems or isolated organs. A major advantage of in vitro studies is that it is possible to test a broad range of high drug concentrations that cannot be reached in vivo. While we do not prohibit publication of assessment of only high concentrations of active compounds in vitro, we discourage observations where few, or no, possibilities exist for in vivo application or where the concentrations are inappropriate for further pharmaceutical development and drug discovery (Heinrich et al., 2020). The natural product literature contains many examples in which high concentrations have been used to evoke a pharmacological effect that is then associated with claims for a therapeutic use (Butterweck & Nahrstedt, 2012; Gertsch, 2009). However, it has been argued that the molecular structures of natural compounds facilitate interaction with proteins and so, at high concentrations, are likely to produce unwanted effects (Gertsch, 2009). Indeed, testing compounds targeting specific proteins (e.g., receptors and enzymes) at concentrations that exceed by 10-fold the IC50 or Ki for the molecular target increases the possibility of introducing off-target actions (Liston & Davis, 2017; Smith & Houghton, 2013). Unfortunately, there are no accepted limits on concentration beyond the cut-off concentrations commonly used in the pharmaceutical industry (EC50 < 10 μM) (Gertsch, 2009). Compound screening assays for hit discovery are typically run at 1- to 10-μM compound concentration (Hughes et al., 2011). A recent guide authored by editors of journals specializing in natural product research suggests that concentrations higher than 30–50 μM should not be used (Heinrich et al., 2020). For the evaluation of certain pharmacological activities, some suggestions are reported elsewhere. For example, antimicrobial activity is believed to be meaningful at concentrations below 25 μM (Cos, Vlietinck, Berghe, & Maes, 2006). For antiproliferative/cytotoxic studies, it is recommended that compounds have selectivity and are not ‘anti-life’ (Heinrich et al., 2020). BJP has elected not to have a policy restricting concentration other than that the concentrations tested are achievable in vivo without causing unwanted biological effects (off or on target). In studies assessing antiproliferative effects in the context of development of treatments to limit carcinogenesis, a comparison between the compound effect on tumours versus healthy cells is recommended. If the aim of the study is to identify the constituent responsible for effects of particular dietary approaches, the effects of the compound must occur at concentrations commensurate with those achieved from dietary consumption. In general, it may be wise to ask the following question: ‘can the concentrations used in vitro be present in the blood and tissue/cellular target after the administration of a therapeutic dose of the compound under investigation’? The answer to this simple question requires knowledge of the pharmacokinetic profile of such a compound. An interesting study that investigated the in vivo relevance of a concentration used in vitro was recently published in BJP (Yeo, Fenwick, Barnes, Lin, & Donnelly, 2017). The authors found that the dietary polyphenol isorhapontigenin inhibited IL-6 release from airway epithelial cells. On the basis of pharmacokinetic data obtained in rats in vivo, the authors postulated that the concentration used in vitro can be reached following oral dosing of isorhapontigenin and that at least 30% inhibition of IL-6 release can be attained through a single oral dose of isorhapontigenin (Yeo et al., 2017). It is important to note that the compound under investigation might be a pro-drug—or, more generally, may be metabolized in vivo before the absorption phase, and hence, the results obtained in vitro could be misleading. This is common for natural products since many plant compounds exist in nature as glycosides, which are generally deglycosylated by intestinal microbiota, making the non-sugar portion (the aglycone) available for absorption. Extrapolation of dose from animal experiments to the human situation can be difficult (Nair & Jacob, 2016). In addition to body weight, a number of variables should be considered, including body surface area, pharmacokinetic parameters (clearance and volume of distribution) and interspecies differences in the pharmacodynamics. A comprehensive analysis of the principles of interspecies dose extrapolation is beyond the scope of this editorial. However, we refer readers to a paper published several years ago in BJP (Sharma & McNeill, 2009). Conversely, resveratrol is an example of a compound where dose ranges of activity in research studies do not match dietary approaches. In an article published in Nature some years ago, resveratrol improved survival of mice on a high-calorie diet (Baur et al., 2006), and this led to speculation that some of the benefits of red wine relate to such an effect. However, extrapolation of the dose used in animals in this study, using allometric scaling (Nair & Jacob, 2016), which considers body weight and surface area, leads to the conclusion that a person of average weight would have to drink 55 bottles of wine per day. In summary, dose ranges in animal experiments must be relevant from a preventive or therapeutic viewpoint and evidence supporting this must be provided. In general, multiple-dose testing is recommended (in the context of compliance with the 3Rs): Single doses are acceptable only in complex pharmacological models. In addition to traditional oral/parenteral dosing methods, compounds under investigation can be incorporated into the diet. An example is a study, published in BJP, that assessed the effects of dietary supplementation for 2 months with the plant product quinic acid on glucose metabolism (Heikkilä et al., 2019). In order to determine how much compound needs to be into the it is to of including the dose that the to to the weight and of the animal The of compounds into the diet has a the study to the relevance of a specific natural compound in the context of a diet. The of drug administration is also the clinical that treatments can only be following and the of administration of compounds under study should be unless the authors are a possible for For in which the is to than compounds should be given after administration of the Many new compounds in clinical they are not (Hughes et al., 2011). at the the of is as important as the of both for of a clinical and for a natural compound as a BJP is a pharmacological and we do not ask for in vivo we authors to provide on the of the product under For example, if a study is in vitro, some of in vitro such as cell or are For in vivo studies, of potential could include for example, of or cardiac For pharmacological activities, such as antiproliferative/cytotoxic activity, the authors should that extracts or compounds have selectivity of action on cells and are not ‘anti-life’ drugs (Heinrich et al., 2020). a comparison should be provided of the effect between and healthy cells the effect is at high authors are also to the literature to if has already been reported. For example, et al. found that the plant cannabinoid at a which was more than than the dose reported in the the BJP has a of to and A major initiative has been the publication and of related to and analysis of The has published on this the for submissions to the et al., et al., 2018). issues in these include a for and in evidence of sample size and a for minimum of to to For all studies animal tissues or authors must issues raised in the BJP and the of for et al., which on research using animals for BJP authors. are to BJP on data and et al., 2017; et al., as well as on the and the of and et al., 2018). these are available on the BJP BJP requires to be considered as a for all et al. 2018). BJP editors that all experiments (in vitro, in vivo and should include both unless there is a specific not to do This the major that authors should consider before on natural products to BJP (see Table It also some common in the pharmacological research on natural products, which could be if experiments have and BJP is a in the pharmacological in which important new are and is a major of for showing the effect of natural products, without a substantial investigation into the mode of action, are not limited to of data or that similar pharmacological activities of similar chemical compounds are generally not for BJP. The editors of BJP the of opportunity for therapeutic that from the natural and are to publish papers on the pharmacology of natural products that of the of both and or that identify potential new therapeutic

Related Papers

• → Natural head posture--considerations of reproducibility(2002)69 cited
• → Short- and long-term reproducibility of radioisotopic examination of gastric emptying(1990)20 cited
• The reproducibility of reported height and body weight in repeated questionnaire surveys.(1995)
• The reproducibility of scores from three video formats.(1988)
• Comparisons of probing depth measurements with Florida probe and conventional periodontal probe(2005)