1. Introduction
The gastrointestinal (GI) tract represents a major interface between the human body and the external environment, functioning as both a vessel for the digestion and absorption of dietary nutrients as well as acting as a major chemosensory organ [
1]. Gut chemosensory mechanisms detect a diverse array of compounds present in the gut lumen and elicit changes to both local and whole-body homeostasis in response to the components detected. These gut chemosensory responses are sensitive to both type and relative location of stimulus, as occurs for example when exposure of gastric tissue to taste modalities stimulate the release of the hunger promoting hormone ghrelin and increases food consumption [
2,
3], whereas the presence of these same tastes in the ileum can stimulate the release of the hunger suppressing hormone glucagon-like peptide-1 (GLP-1) and decrease food consumption [
3,
4]. These distinct response profiles suggest that characterization of not only chemosensory systems present, but also their localization is important in understanding gut chemosensing.
The gut chemosensory system is comprised of various receptors and transporters [
5,
6,
7], with growing interest in flavor and taste modalities such as bitterness [
8]. In humans, bitter taste is detected by a family of 25 bitter taste receptors (TAS2Rs) that respond to both dietary and non-dietary compounds [
9]. Lingual bitter taste is a sensation that potentially signals the presence of a toxin and is more pronounced in infants and children [
10,
11]. Extra-lingual gut TAS2Rs are present in the gut mucosa including within the specialized chemosensing enteroendocrine cells [
12], that represent less than 1% of the total gut epithelial cell population and are broadly defined by the peptide hormones they contain [
13,
14]. These peptide hormones include, the appetite regulating cholecystokinin (CCK), GLP-1, peptide tyrosine tyrosine (PYY) and ghrelin, the intestinal motility modulating neurotensin and serotonin, and the gut secretagogues, gastrin and secretin [
15,
16]. The presence of extra-lingual TAS2Rs in enteroendocrine cells infers bitter taste modalities may be gut active and affect numerous physiological processes, ranging from the regulation of ingestive behavior [
3,
17,
18,
19] and gut motility [
20] to adverse GI responses to pharmaceutical products [
21] and chemical absorption [
22]. Additionally, single-cell sequencing of human intestinal tissue has also demonstrated the presence of specific TAS2Rs in intestinal goblet and Paneth cells, and rodent studies have shown TAS2R presence in Tuft cells [
23,
24], suggesting a potential role in the modulation of gut secretion, mucin homeostasis, and immunity [
25,
26,
27].
Previous studies have detected the presence of TAS2Rs within the human gastrointestinal tract [
28,
29,
30,
31,
32], and a recent review indicates likely regional variation in human TAS2Rs [
32]. However, due to challenges mapping the lowly expressed TAS2R transcripts, it is still unclear the degree to which TAS2Rs vary by region of the GI tract, whether all or a subset of TAS2Rs are present in specific regions of the GI tract mucosa, and if there is conserved or differential TAS2R expression between individuals.
Here we conduct a comprehensive study of human intestinal TAS2R expression, examining biopsy samples from 28 healthy male and female adults taken from 10 regions throughout the human GI tract. We show that TAS2R transcripts are present, that specific TAS2R transcripts vary between gut regions and highlight a subset of individuals who exhibit a broader TAS2R transcriptional profile.
3. Discussion
Gastrointestinal tract bitter chemosensing relates to multiple aspects of gut physiology [
3,
33,
34,
35,
36], it may affect gut-brain signaling and regulate whole body homeostasis [
37,
38]. Preceding studies have shown the presence of human TAS2Rs in biopsy samples from the stomach [
39], small intestine[
31] and colon [
29], providing evidence for roles in intestinal function, immunity and feeding behavior. Given the broad scope of the current study and the use of a sensitive detection method that allows for the assessment of lowly transcribed receptors, the results presented here extend our knowledge of human GI TAS2R expression. We detect transcripts for most TAS2Rs in all regions of the GI tract, show novel expression patterns reflecting differential TAS2Rs expression based on GI region and chromosomal location, and highlight for the first time 2 archetypes of GI bitter sensing.
TASR2s are expressed throughout the GI tract.
In general agreement with prior work examining GI expression of human TAS2Rs, we show gastric, small intestinal and colonic expression, however several differences in TAS2R expression profile present that are worthy of discussion. Expression of TAS2R7, TAS2R10, TAS2R14, TAS2R43 and TAS2R46 was shown in the Gastric Fundus by Liszt et al. 2017 [
39] using q-PCR assessment of 2 human biopsy samples. As compared to Liszt et al. 2017 [
39] we show expression of TAS2R10, 14 and detect the presence of the TAS2R43 superfamily in the Gastric Fundus. Additionally, we observe variable TAS2R7 expression that is not reported in Liszt et al and may reflect the limited sample number taken during their work. Le neve et al. 2010 [
31], using pooled RNA samples from 4 male donors, showed TAS2R14 but not TAS2R7 expression in human Duodenum and Jejunum samples. Our results are in line with these, showing TAS2R14 present in the small intestinal samples, whereas, TAS2R7 was not detected in samples taken from the majority of participants. Kaji et al. 2009 [
30] , Latorre et al. 2016 [
12] and Rozengurt et al. 2006 [
29] reported the presence of TAS2Rs in the large bowel. Kaji et al. 2009 [
30] showed TAS2R38 to have consistent transcription across the Ascending Colon, Sigmoidal Colon and Rectum, and showed weak banding for TAS2R1 and 4 [
30]. Latorre et al. 2016 [
12], observed TAS2R38 in enteroendocrine cells of the lower Colon that they confirm with antibody staining. Rozengurt et al. 2006 [
29] showed 18 out of the 25 human TAS2Rs were present in a pooled sample of commercially available human Colon RNA. The data presented here is broadly in agreement with the previous published assessments of large bowel TAS2R expression. We find evidence for the presence of TAS2R4 and TAS2R38, but not TAS2R1 in samples from most participants. Of the 7 TAS2Rs that Rozengurt et al. [
29] did not detect, we also do not identify TAS2R 1, 7, 8, 9, 16 and 41 in the majority of colonic samples. Conversely, we show that TAS2R14, undetected in Rozengurt’s study, was the most abundant in our data. The data presented here is also in general agreement with consensus data of human GI tract TAS2R expression reported in a recent review [
32].
Gastrointestinal TAS2Rs exhibit regional variation.
TAS2R transcription showed significant variation along the length of the human GI tract. Generally, TAS2Rs exhibit two gross patterns of transcription with the highest transcript being in the Gastric Fundus, Gastric Body and Terminal Ileum for TAS2R3, 4, 5, 13, 14, 19, 20, 43 and 50, whereas other TAS2Rs were elevated in the small or large intestinal regions (TAS2R38, 40, 41 and 60). Overall, the Gastric Fundus and the Terminal Ileum recorded the highest TAS2R counts, suggesting enhanced chemosensing capacity in these gut regions that may reflect likely exposure to dietary and microbiome derived compounds respectively. Interestingly, TAS2R38 was most highly transcribed in the terminal Ileum, is a known receptor for amine moieties containing molecules (such as acetythiourea and caprolactum) [
9] and its allelic haplotype is implicated in respiratory response to bacterial infection [
40] . This, combined with the TAS2R38’s specificity towards bacterial quorum sensing compounds [
41], suggests TAS2R38 may have a potential role in bacterial overgrowth sensing in the terminal Ileum. A degree of SI regionality, particularly in the TI, was also observed for TAS2R41 and 60. Small intestinal enteroendocrine cells contain appetite suppressing and insulin sensitizing hormones, such as CCK, GLP-1 and PYY [
42], with the activation gastrointestinal of TAS2Rs implicated in the reduction of food intake and enhanced glucose control [
3,
38]. The delivery of a TAS2R41 or TAS2R60 specific agonist could potentially activate only small intestinal TAS2Rs, triggering the release of appetite suppressing peptide hormones without the unwanted activation of gastric or colonic enteroendocrine cells [
43,
44]. To date, TASR2R41 has only two known agonist, the antibiotic chloramphenicol and the artificial sweetener sucralose [
45,
46], neither of which are specific to TAS2R41 or likely to be the endogenous or environmental ligands, TAS2R60 is an currently an orphan receptor.
Gastrointestinal TAS2Rs exhibit inter-individual variation.
We introduce the concept of gut super tasters as evidenced by a more complete repertoire of TAS2Rs throughout their GI tract. This complement of TAS2Rs is present in all GI samples taken from these individuals and suggests a conserved regulatory mechanism that is unlikely to be directly affected by exposure to specific dietary/bacterial compound(s). This conclusion is further supported as one participant, who volunteered samples from both a gastroscopy and colonoscopy, exhibited an enhanced TAS2R repertoire along the entire length of the GI tract. Importantly, an expanded GI TAS2R repertoire may confer a broader chemosensory capacity to these individuals that may modulate their GI responses to specific compounds, be they dietary, bacterial or endogenous in nature. For example, TAS2R16 was detected only in the gut super tasters and is the sole receptor for the detection of natural compounds arbutin and amygdalin D [
9], which are present in the leaf of bearberry plants and stone fruit kernels respectively. The presence of GI TAS2R16 may allow for a more pronounced gastrointestinal response to these plant toxins and hence may afford additional protection from accidental poisoning when consuming the associated food. Additionally Salicin, the willow bark derived anti-inflammatory agent [
47] , is also a TAS2R16 specific agonist that may trigger GI side effects in some people [
48]. It is possible that these reported side effects are, in part, mediated by the actions of GI TAS2R16 in these individuals, and that side effects for other drugs are in part dictated by the presence of specific TAS2Rs. We believe this is the first evidence suggesting the existence of gut super tasters and has ramifications for interventions targeting GI TAS2Rs, and for the understanding of gastrointestinal side effect to pharmaceuticals and bitter bioactives. Unlike the oral super tasters, “gut super tasters” are not defined based on their specific genetic polymorphism, but on a general increase in the spectrum of TAS2Rs present with the gut. It is possible that along with a general increase in the spectrum of TAS2Rs present, there may be genetic polyphorphisms and/or changes in the absolute number of enteroendocrine cells. Indeed, recent studies report changes in both enteroendocrine cell number in response to diet and highlight the role of haplotype variation in oral bitter taste perception [
49,
50].
Expression levels vs chromosome location and gene relatedness
Human TAS2Rs are located on 3 cytogentic locations, 5p15, 7q31-35 and 12p13 that contain a single gene, a cluster of 9 genes and a cluster of 15 genes respectively. As previously reported for rodent TAS2Rs [
51], TAS2Rs from the same chromosomal location exhibited similar patterns of transcription along the human GI tract suggesting conserved cis-regulatory elements between closely located genes. Although relative GI patterns of transcription were conserved within a given chromosomal location, absolute transcription number was not. It is possible that there is copy number variation, deletions of regulatory factors that affects the gene transcription, or the presence of gene choice elements such as those that effectively turn on or off transcription of adjacent odorant receptor genes [
52]. TAS2R41 and 60 both reside on chromosomal region 7q35, show close phylogenetic relationship and are the only TAS2R genes that have significantly elevated small intestine transcript count. Again, this suggests that in addition to genetic and locational similarities, they may also have conserved regulatory elements and their transcription maybe regulated together.
Strengths and limitations
There are several limitations to the current work, including the characterization of transcript rather than protein expression, the inability to localize TAS2R expression within a specific cell type, and the assessment of transcript without additional haplotype analysis. These limitations were considered during the study design and reflect decisions made to increase participant recruitment rates by not assessing genetic information, and to ensure optimal detection of lowly expressed TAS2Rs by extracting RNA from the entire biopsy sample. Further to this, the characterization of TAS2R proteins was considered impractical due to the current lack of effective TAS2R antibodies [
53]. Strengths of this study include the relatively large number of participants, the strict criteria for participant inclusion and the choice of the highly sensitive NanoString detection technology that allows for detailed resolution of lowly transcribed genes without pooling clinical samples [
54].
4. Materials and Methods
Figure 4.
Schematic overview of methods: Participants undergoing routine Gastroscopy and/or Colonoscopy were recruited and volunteered 10 biopsies per procedure taken in biopsy pairs. Dual biopsies were processed together with total RNA extracted. TAS2R analysis of total RNA samples was conducted using Nanosting™ direct RNA counting and analyzed for gastrointestinal expression, regional variation, and inter-individual variation.
Figure 4.
Schematic overview of methods: Participants undergoing routine Gastroscopy and/or Colonoscopy were recruited and volunteered 10 biopsies per procedure taken in biopsy pairs. Dual biopsies were processed together with total RNA extracted. TAS2R analysis of total RNA samples was conducted using Nanosting™ direct RNA counting and analyzed for gastrointestinal expression, regional variation, and inter-individual variation.
4.1. Ethics and Recruitment
Ethical approval was obtained from the New Zealand Health and Disability Ethics (HEDC) committee prior to the commencement of the study (NTY/11/08/077). This research was conducted in alignment with the Declaration of Helsinki and all participants gave informed consent before taking part in this study, Participants were recruited from routine endoscopy lists at regional hospitals of the Waitemata District Health Board (Auckland, New Zealand). Participants were approached prior to undergoing either a colonoscopy or a gastroscopy, a medical history was taken, suitability for the trial was determined and informed consent was obtained. Inclusion criteria for this study was: male or female, 18-65 years of age, normal gross gastrointestinal tract anatomy, generally healthy as ascertained by self-report, undergoing screening gastroscopy or colonoscopy, and provision of written informed consent. Exclusion criteria for this study was: obese as defined by BMI >30kg/m2, previous GI tract resection (including gastric banding), diagnosed diabetes mellitus (type 1 or 2), diagnosed medical conditions known to affect gut function including but not limited to malignant or inflammatory diseases of the gut such as: inflammatory bowel disease (Crohns, colitis), ischemic bowel, or coeliac disease. Current or recent use of the following medications: non-steroidal anti-inflammatory drugs (NSAIDs), proton-pump inhibitors, glucocorticoids, immunomodulators, anticoagulants, weight loss medication, and alcohol or drug dependency. Extreme food restriction and/or excessive physical activity as seen in anorexia nervosa, bulimia or similar eating disorders that alter body composition in a short space of time, and/or purging (vomiting/using laxatives). Unwilling/unable to comply with study protocol. Participating in another clinical intervention trial.
4.2. Biopsy Tissue Sampling
Table 3.
Location of Biopsy Sampling.
Table 3.
Location of Biopsy Sampling.
Sample Number |
Procedure |
Sample Name |
Location Notes |
1 |
Gastroscopy |
Gastric Fundus |
|
2 |
Gastroscopy |
Gastric Body |
Approximately midpoint along the greater curvature |
3 |
Gastroscopy |
Gastric Antrum |
|
4 |
Gastroscopy |
D2 Duodenum |
|
5 |
Gastroscopy |
Proximal Jejunum |
Or as close to the Duodenojejunal Flexure as possible |
6 |
Colonoscopy |
Terminal Ileum |
Proximal to the ileocecal sphincter |
7 |
Colonoscopy |
Ascending Colon |
|
8 |
Colonoscopy |
Transverse Colon |
|
9 |
Colonoscopy |
Sigmoid Colon |
|
10 |
Colonoscopy |
Rectum |
|
Dual 2-3 mm intestinal biopsy samples were taken from gastric fundus, gastric body, gastric antrum, duodenum (2nd part (D2) and either the proximal jejunum or 4th part (D4) of the duodenum (referred to as proximal jejunum sample), for participants undergoing gastroscopy, and rectum, sigmoidal colon, transverse colon, ascending colon and terminal ileum for participants undergoing colonoscopy (
Supplementary Figure S1). Samples were rinsed in RNase free water (Life Technologies, Carlsbad, CA, USA, cat # 10977-023), and immediately immersed in RNAlater® (Life Technologies, cat # AM7021). Samples were kept at 4
oC for 24h then frozen at -80
oC until analysis, as described in the manufactures protocol.
4.3. NanoString Bitter Taste Receptor Analysis
Dual biopsy samples were processed together using a TissueLyser II (Qiagen, Limburg, Netherland, cat #: 85300), total RNA was extracted using QIAzol Lysis Reagent (Qiagen, Cat #: 79306) and RNA extraction kit (Qiagen, cat #: 217004) as described in the manufactures’ protocol. Samples were analysed for mRNA transcription of human TAS2Rs by NanoString® (NanoString® Technologies, Inc. Seattle, WA). RNA Probes were designed by NanoString® Technologies Inc (
Supplementary Figure S2). NanoString® experiments were conducted by New Zealand Genomics Ltd (Dunedin, New Zealand) and analysed at The New Zealand Institute of Plant and Food Research Ltd (Auckland, New Zealand) as described in the manufacture’s protocol (see
Supplementary Table S1 for probe design). Sequence similarity of 5 TAS2Rs (Ta2R43, 44, 45, 46 and 47) within the probe target areas made the design of specific probes impractical and 1 probe (TAS2R43) detected all of the members of this super family with greater than 93% efficiency. Briefly, RNA was quantified using a NanoDrop 2000 UV-Vis spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). 600ng per sample was added to the Nanostring® probe hybridisation reaction and incubated for 18h at 64
oC. Hybridized probes were then purified using the nCounter® Prep Station (NanoString Technologies, Inc.). Samples were washed to remove excess capture and reporter probes then immobilized on a streptavidin-coated cartridge to remove transcript-specific ternary complexes. Data collection was carried out on the nCounter® Digital Analyser (NanoString® Technologies, Inc.) counting individual fluorescent barcodes to quantify target RNA molecules present in each sample. Analysis was conducted as per the manufactures’ recommendations. Briefly, raw Count data from the nCounter was normalized relative to a standard curve constructed using spike in exogenous control samples, then subsequently to a standard curve generated using the geometric mean of 6 reference genes B2M (NM_004048.2), G6PD (NM_000402.2) GAPDH (NM_002046.3), YWHAZ (NM_003406.2), TUBA1A (NM_006009.2) POLR2A (NM_000937.2). The background hybridization signal was determined using the spike in negative controls. All non-background RNAs underwent a background correction by subtraction of the mean counts of the negative controls +2 standard deviations. Final RNA counts that were below 50 were classified as being minimally detectable.
4.4. Statistical Analysis
For statistical analysis, normalized mRNA count data (as described above) were log transformed. Analysis of variance testing (ANOVA) was performed to determine if significant differences in RNA count existed between gut sites and between participants. When significant differences did occur, post-hoc analysis was conducted using Tukeys HSD to determine where these differences occurred.
For the investigation into the clustering of participants the gut sites were split into two groups, the upper Gastroscopy sites and lower Colonoscopy gut sites. Similar to all other analyses the response variable (gene expression) underwent log transformation. This analysis excluded participant 6 due sample BS31’s failure at QA.
Multiple methods for cluster analysis were conducted to determine presence and number of clusters for TAS2R RNA count. For cluster analysis the following procedures were employed: 1. ANOVA analysis to test whether there is any evidence of difference between participants. 2. Hierarchical clustering dendrograms using ward.d2 clustering criteria (24). 3. A plot of within group sum of squares by number of clusters by K-means partitioning. 4. NbClust (25) package within R, for both ward.d2 and k means with 26 criteria. For each method the function applies 26 criteria for determining the number of clusters, it then takes the best number of clusters as the one with the majority of the criteria suggesting that it is the best partition. 5. Genes identified from previous plot of expression vs participant, any with significant ANOVA p-values and a consensus of two clusters were further investigated to see how participants were grouped. 6. Several genes were found to have the same grouping of participants in their level of transcription. With the participants who are in both upper and lower gut-sites being present in the same groups for both regions. The selected genes were then combined, and a principal component analysis performed (PCA). From the PCA we produced a cluster-plot for the participants.
All the analyses other than those specifically noted were conducted in R: A language and environment for statistical computing. A P value of < 0.05 was designated as significant.