Previous studies have demonstrated that the TAS2R43 bitter taste receptor is a key mediator of aristolochic acid perception. Functional forms of the receptor exhibit detectable responses at concentrations of ~10 nanomolar in heterologous expression assays, and some individuals reliably perceive aristolochic acid at concentrations as low as 40 nanomolar [15
]. It has also been demonstrated that ability to perceive aristolochic acid is influenced by common functional mutations in TAS2R43 including two tightly linked amino acid variants, C104G/S35W and A635G/H212R (rs68157013 and rs71443637), and a whole-gene deletion, TAS2R43-Δ. While TAS2R43-W35/H212 exhibits strong response to aristolochic acid in heterologous expression assays and is associated with low threshold detection (i.e.
, high taste sensitivity), TAS2R43-S35/R212 and TAS2R43-Δ exhibit weak or no response and are associated with high threshold (i.e.
, low sensitivity). These findings have suggested that variation in TAS2R43 might affect susceptibility to BEN, with TAS2R43-W35/H212 conferring protection by signaling aristolochic acid contamination in food.
SNP genotyping and copy-number determination revealed that the three major functional variants of TAS2R43 (TAS2R43-W35/H212, -S35/R212 and –Δ) were all common in our sample, with frequencies of 0.17, 0.36, and 0.47 respectively. No recombinant variants of TAS2R43-W35/H212 and -S35/R212 (e.g., -W35/R212) were observed, such that W35S and H212R were in perfect linkage disequilibrium (LD) (r2
1). These patterns are consistent with those reported by Roudnitzky et al. [19
], who observed frequencies of 0.15, 0.35, and 0.50 in a similar panel of European subjects, with tight linkage (r2
1) between W35S and H212R. Overall, 2.7% (n
5) of subjects were homozygous for the TAS2R43-W35/H212 allele, 29.4% (n
55) were heterozygous, and 67.9% (n
127) were non-carriers (Figure ). Again, these values are comparable to those observed by Roudnitzky et al. [19
] (3.3%, 23.3%, and 73.3%). Thus, levels and patterns of variation in our sample appear to be typical for European populations. However, they are not necessarily typical for other populations. Pilot data from the 1000 Genomes Project (1000GP), a consortium-based effort to sequence more than 2000 complete human genomes with worldwide representation, indicate that TAS2R43-S35/R212 is present at frequencies <3% in Asian populations [21
Genetic data. AOccurrence of TAS2R43-W35/H212, -S35/R212 and Δ alleles in subjects. Counts are given, with frequencies shown in parentheses. B Genotype frequencies in subjects.
Association tests detected a significant relationship between TAS2R43-W35/H212 carrier status and BEN; however, counter to expectation, the correlation was positive, not negative as would be predicted if perception of aristolochic acid leads to avoidance (Figure ). Fisher’s exact test yielded a P
-value of 0.020, with the case population being enriched with TAS2R43-W35/H212. Tests using linear regression indicated the presence of an effect with an estimated odds ratio of 1.18 (P=
0.025). Similar P
-values were obtained using linear regression (P
0.11) and χ2
0.025). Thus, significance levels were not sensitive to the analytical model. These findings are at odds with the prediction that aristolochic acid-sensitive TAS2R43 alleles provide an effective safeguard against overexposure, and raise the paradoxical possibility that they instead confer susceptibility.
Figure 3 Association data and tests. A Frequencies of genotypes with respect to functionally divergent allele categories N (Non-responsive to aristolochic acid; W35/H212) and R (Responsive to aristolochic acid; S35/R212 and Δ). B Results of regression, (more ...)
Evidence that aristolochic acid-sensitive alleles confer increased susceptibility to BEN is striking and counterintuitive. Bitter perception has long been hypothesized to provide protection against toxins in the environment, particularly plant toxins, by enabling detection and avoidance [22
]. This hypothesis is supported by much circumstantial evidence: bitter tastes are generally aversive, plant-derived toxins are over-represented among known bitter tastants, bitter taste receptors respond to many well-known toxins, and plants containing these toxins are perceived as bitter [9
]. Thus, the clear prediction under the toxin-detector hypothesis is that TAS2R43-W35/H212, which has been shown to be associated with low taste thresholds for aristolochic acid, should provide a warning signal providing protection against BEN or, if aristolochic acid concentrations are too low to perceive, have no effect on susceptibility. Our finding that TAS2R43-W35 is positively, rather than negatively, associated with BEN is strongly at odds with this prediction.
One explanation for our findings could be that, contrary to expectations, the bitter taste of aristolochic acid is attractive to those who perceive it, rendering them susceptible to disease-causing levels of exposure. This explanation conflicts with predictions under the toxin-detector hypothesis; however, it is consistent with psychophysical studies, which have revealed that while intense bitter sensations are universally aversive, mild sensations can evoke positive responses [23
]. Positive aspects of bitterness are evident in everyday experience; many foods, such as beer and salad greens have bitter, yet overall pleasurable, taste characteristics. This explanation is consistent with evidence that BEN arises as the result of chronic low-level exposures, which might be perceived as pleasantly bitter, as opposed to acute high-level exposures, which would likely be aversive.
A second, more speculative, explanation for our results is suggested by reports that bitter taste receptors and other major components of the taste transduction pathway are expressed in the mammalian gut, as well as on the tongue [24
]. In mice and rats, Tas2R transcripts are found in gastric and duodenal tissues, and cells derived from intestinal tissue show dosage-dependent responses to bitter tastants [25
]. The native role of TAS2R-mediated signaling in the gut remains unclear, but signals initiated by these receptors can induce the release of signaling peptides from enteroendocrine cells [25
]. Further, some evidence suggests that post-absorptive metabolic activation of aristolochic acid is crucial to its toxicity [26
]. If the response of gut-expressed TAS2R43
to aristolochic acid enhances either absorption or metabolism, then genetic variation in TAS2R43
might shape susceptibility to BEN.
BEN’s distinctive pathology is not unique. Similar patterns of kidney damage and increased rates of urothelial cancer have emerged in subjects overexposed to herbal supplements containing aristolochic acid [2
]. Based on the etiological similarities of these diseases, it has been suggested that they be aggregated under an umbrella term, aristolochic acid nephropathy (AAN). To date, efforts to find genetic variants accounting for variable susceptibility to AAN have been relatively unsuccessful. For instance, variants in several genes (NQO1
, and NAT2
) found previously to be associated with occupationally-induced renal and bladder cancers were not associated with AAN diagnosis [17
]. Our findings recommend that variation in TAS2R43
be considered a candidate factor in overall AAN susceptibility. While taste sensitivity alone seems unlikely to account for variable exposure aristolochic acid via nutritional supplements, which are often encapsulated to disguise taste, TAS2R43 variation might exert effects through activity in the gut.