Metabolomics reveals diet-derived plant polyphenols accumulate in physiological bone

Abstract

Plant-derived secondary metabolites consumed in the diet, especially polyphenolic compounds, are known to have a range of positive health effects. They are present in circulation after ingestion and absorption and can be sequestered into cells within particular organs, but have rarely been investigated systematically in osteological tissues. However, a small number of polyphenols and similar molecules are known to bind to bone. For example alizarin, a plant derived anthraquinone and tetracycline (a naturally occurring antibiotic), are both absorbed into bone from circulation during bone formation and are used to monitor mineralization in osteological studies. Both molecules have also been identified serendipitously in archaeological human bones derived from natural sources in the diet. Whether an analogous mechanism of sequestration extends to additional diet-derived plant-polyphenols has not previously been systematically studied. We investigated whether a range of diet-derived polyphenol-like compounds bind to bone using untargeted metabolomics applied to the analysis of bone extracts from pigs fed an acorn-based diet. We analysed the diet which was rich in ellagitannins, extracts from the pig bones and surrounding tissue, post-mortem. We found direct evidence of multiple polyphenolic compounds in these extracts and matched them to the diet. We also showed that these compounds were present in the bone but not surrounding tissues. We also provide data showing that a range of polyphenolic compounds bind to hydroxyapatite in vitro. The evidence for polyphenol sequestration into physiological bone, and the range and specificity of polyphenols in human and animal diets, raises intriguing questions about potential effects on bone formation and bone health. Further studies are needed to determine the stability of the sequestered molecules post-mortem but there is also potential for (palaeo)dietary reconstruction and forensic applications.

Figure 1

Polyphenols identified in the acorn supplement and measurements for the identification of urolithin A in pig bone extracts. (a) Bar graph showing magnitude of the ion count for individual polyphenols identified in the acorn extracts (error bars St. dev. n = 9/group). (b) Average peak areas for plant secondary metabolites identified from acorn extracts (St. dev. n = 9/compound). (c) Chromatographic retention time for Urolithin A in the authentic standard and Sardinian bone 6070116 is 7.42 and 7.43 mins respectively, a difference of 0.6 seconds. (d) The accurate m/z measurement of Urolithin A in the authentic standard and Sardinian bone provided a difference of 0.85 ppm. (e) Comparison of the relative abundance of the M, M + 1 and M + 2 isotopes for Urolithin A in the authentic standard compared to Sardinian bone showed a 99% similarity. (f) Comparison of the product ion spectra from the collision induced dissociation (CID) of urolithin A (precursor m/z 227.03) in the Sardinian extract (top) and product ion spectra from the analysis of the urolithin A standard (bottom).

Figure 2

Identification of plant-derived polyphenols in animal bone. (a) Relative abundance (chromatographic peak area) for urolithin A and urolithin B in extracts from each of the Sardinian bones (mean and St. dev. n = 5). (b) Table showing the matching criteria for the identification of diet-derived compounds in bone extracts. (c) Structures of the five plant-derived compounds identified in pig bones from the animal feeding study. (d) External calibration curve showing linearity for the urolithin A peak area response to concentration over the range 10 ng/mL to 0.08 ng/mL. The peak areas generated for urolithin A in each of the Sardinian bone extracts are also plotted and labelled (red triangles) on the line of best fit.

Figure 3

Identification of plant-derived polyphenols in bones and bone model. (a) Relative ion abundances for the 5 metabolites identified from our in house ‘polyphenol database’ across the 5 Sardinian bones extracts (error bars st. dev; n = 3). (b) PCA scores plot shows samples cluster into dietary groups based on their compound feature composition. (c) Table showing the calculated amount of urolithin A in each of the five Sardinian pig bone samples (NF-not found) (mean values n = 5). (d) A list of percentage binding to hydroxyapatite (HAP), for polyphenolic compounds chosen to demonstrate a range of molecular weights and structures. Further details along with compound structures provided in an expanded table in SI (Supplementary Fig. 6 and Table 3).