Introduction

Trace amine-associated receptors (TAARs) are a novel class of G protein-coupled receptors. Their name was derived from the observations that some receptors of this class are activated by amines, which are present in mammalian tissues at very low concentrations (in the order of 0.1–10 nM) and are therefore known as “trace amines” (Zucchi et al. 2006). The latter include para-tyramine, meta-tyramine, tryptamine, β-phenylethylamine, para-octopamine, and meta-octopamine (Berry 2004). However, TAARs include nine different subtypes, and most of them have been considered as orphan receptors because they do not appear to interact with trace amines.

In 2006, Liberles and Buck reported that in mouse, all TAAR subtypes except TAAR1 are expressed in a small subpopulation of olfactory neurons, suggesting that these neurons use TAARs rather than odorant receptors to detect chemosensory stimuli. Like odorant receptors, individual mouse TAARs are expressed in unique subsets of neurons dispersed in epithelium, and there is indirect evidence that they may be coupled to Gαolf proteins, the G protein subtype to which odorant receptors couple. In heterologous cell models, most murine TAAR (mTAAR) subtypes were shown to be activated by several volatile amines. Interestingly, at least three mTAARs interacted with endogenous amines that may play a role in behavioral responses: mTAAR4 was activated β-phenylethylamine, whose elevation in urine is related to stress responses in both rodents and humans, while both mTAAR3 and mTAAR5 detect amines that are enriched in male versus female mouse urine, namely isoamylamine and trimethylamine, respectively. The former is assumed to be a pheromone since it accelerates the onset of puberty in female mice (Liberles and Buck 2006).

TAAR expression has also been detected in the olfactory epithelium of teleost fish (Hashigushi and Nishida 2007) and Xenopous laevis (Gliem et al. 2009), but no data is available in human. Therefore in the present research, we investigated whether TAAR genes are expressed in cells of the human olfactory epithelium, using biopsies of human olfactory mucosa obtained from healthy patients undergoing routine surgery on the septum for therapeutic reasons.

Methods

Patients

Our series included 16 patients (12 men and 4 women, aged from 16 to 54 years, with an average age of 31.7 ± 3.2 years) who underwent routine nasal surgery (septoplasty and turbinectomy) from December 2007 to January 2009 at the ENT Division of Pisa University Hospital for nasal septum deviation and turbinate hypertrophy. In each patient, a detailed medical history was recorded, and a complete rhinological evaluation was performed. Clinical data are summarized in Table 1. Four out of 16 subject were smokers, but only one of them had been smoking more than 10 cigarettes a day for over 10 years. No one had been exposed to dust, fumes, or irritants. Five patients were suffering from allergic rhinitis, and one had allergic asthma. Two patients were affected by obstructive sleep apnea syndrome (OSAS). None of the patients showed any evidence of acute inflammatory disease of the upper respiratory tract. None of the patients had a history of abnormal olfactory function, and none had undergone previous nasal or sinus surgery. In each case a high-resolution computed tomography (HRCT) of the maxillofacial region showed no evidence of paranasal sinuses disease.

Table 1 Patient’s data
Full size table

Biopsy Procedure

After obtaining informed consent from each patient, samples were collected from the nasal mucosa during routine nasal surgery. Biopsies were performed under general anesthesia, with the use of giraffe cup forceps (3-mm, 70° upturned, vertical opening cupped giraffe forceps) under endoscopic guidance, in order to minimize tissue injury and avoid olfactory dysfunction. On the whole, 35 biopsy specimens were obtained from our 16 patients. Sixteen specimens were taken on the lateral wall, near the anterior middle turbinate insertion or in its medium part of it; another 16 specimens were taken on the upper portion of the septal wall or in the area where the cartilaginous and bony nasal septum are contiguous, opposite the high middle turbinate, or in the dorsoposterior regions of the septum (Lanza et al. 1993; Leopold et al. 2000). Three control specimens were taken from the inferior turbinate mucosa, which is a region devoid of olfactory neuroepithelium (Escada et al. 2009) (Fig. 1a). Biopsies were obtained from healthy pink mucosa, and their average size was 0.3 × 0.2 × 0.1 cm. Two biopsies from oropharyngeous mucosa of posterior tonsillary pillar were also taken from patients undergoing tonsillectomy.

Fig. 1
figure 1

a Picture of the medial and lateral nasal walls showing the areas in which biopsies were performed. The sites including putative olfactory mucosa are indicated with “S” (septal area) and “T” (turbinate area). Control samples, indicated with “C”, were taken from the inferior turbinate. b OMP expression. Samples were classified as OE positive (black bars) if expression levels fell outside the 99% confidence limits (black line) derived from the control samples (gray bars). White bars represent OE negative samples. See text for further details. c Graphical synthesis of TAARs expression. Bars represent TAAR5 expression, while the table summarizes the expression of the other TAARs

Full size image

Gene Expression

The expression of human TAAR (hTAAR) genes was assessed by absolute quantitative real time PCR using human genomic DNA as an external standard. Nasal mucosa and tonsil biopsies were treated with RNALater Solution (Ambion Inc, USA) immediately after dissection. After overnight preservation at 4°C, they were homogenized in Trizol Reagent (Invitrogen Ltd, UK) using Tissue Ruptor Homogenizer (Qiagen, Germany) with disposable probes. Total RNA was then extracted following Trizol protocol. RNA was incubated with DNase I (Sigma, MO, USA) for 30 min at 37°C and purified with Pure Link Micro to Midi Total RNA Purification System (Invitrogen Ltd, UK). RNA quantification RNA was performed with Qubit Spectrofluorimeter (Invitrogen Ltd, UK) and quality checked by 2100 BioAnalyzer (Agilent Technologies, CA, USA). A total RNA sample from a human lymphocytary pellet was kindly provided by Prof. Mario Petrini (Dept of Hematology, University of Pisa)

Human genomic DNA was purified from a lymphocytary pellet using DNA Easy Mini Kit (Qiagen, Germany) and quantified by Qubit Spectrofluorimeter. One microgram of RNA was retrotranscribed with Quantitect Reverse Transcription Kit (Qiagen, Germany) following manufacturer's protocol. The obtained cDNAs were used for absolute quantitative real time PCR analysis on all the six human TAAR genes (hTAAR1, hTAAR2, hTAAR5, hTAAR6, hTAAR8, and hTAAR9 since hTAAR3, hTAAR4, and hTAAR7 are pseudogenes) on olfactory marker protein (OMP) gene and on hypoxanthine-guanine phosphoribosyl transferase (HPRT) gene, which is a housekeeping gene that was included as an internal control.

Reactions were performed in a total volume of 20 μl containing cDNA equivalent to 50 ng of total RNA, 0.2 μM each oligonucleotide, and 10 μl of iQ SYBR Green Supermix (Bio-Rad, CA, USA). Real time PCR was conducted on an iQ5 Optical System (Bio-Rad, CA, USA) with the following cycle program: 30 s at 95°C, followed by 40 two-step amplification cycles consisting of 10 s denaturation at 95°C and 30 s annealing/extension at 60°C. A final dissociation stage was run to generate a melting curve for verification of amplicon specificity and primer dimer formation. An additional control was performed by 2% agarose gel electrophoresis (Suppl Fig. 1). Specificity was also confirmed by amplicon sequencing from genomic DNA and also from a few samples of cDNA. For each gene, a standard curve was constructed with six 4-fold serial dilutions of human genomic DNA, starting from 10 or 50 ng (depending on expression levels of the gene). Micrograms of human genomic DNA were converted into copy numbers considering C-value, the mass of the haploid human genome (http://www.genomesize.com). All amplification efficiencies were in the optimal range (90–105%). All samples were run in duplicate. Absence of genomic DNA contamination in cDNA samples was confirmed by real time PCR and agarose gel electrophoresis of retrotranscription reactions performed without reverse transcriptase. No signal was detected in any sample (Suppl Fig. 1). Absolute TAARs, OMP, and HPRT cDNA copy numbers were calculated from standard curves. Primers were designed on the basis of coding sequences published on GenBank using Beacon Designer 7 software (Premier Biosoft International, CA, USA). Oligonucleotide sequences are listed in Table 2.

Table 2 Primers used for real time PCR
Full size table

Statistical Analysis

Results are reported as mean ± SEM. Different groups were compared by contingency analysis performed through Fisher's test. Correlations between groups were studied by linear regression. The threshold of statistical significance was set at P < 0.05. GraphPad Prism version 4.1 for Windows (GraphPad Software, CA, USA) was used for data processing and statistical analysis.

Results

Septal and turbinate biopsies were obtained from each patient. In three patients (#13, 14, and 16), samples were also taken from nonolfactive areas in the inferior turbinate. No complication was recorded in any patient, and no patient reported subjective olfactory dysfunction either immediately after surgery or during the 6- to 24-month follow-up period.

The expression of all human TAARs and of the olfactory epithelium marker OMP was determined by absolute quantitative real time PCR, and the results are shown in Table 3. OMP transcripts were detected in all samples, although the number of copies varied over a wide range, namely from 74 to 11,800 cDNA copies per 50 ng RNA (see Fig. 1b). OMP mRNA was also detected in the three inferior turbinate mucosa biopsies. As a matter of fact, OMP mRNA has never been quantitated in human olfactory epithelium, and microarray expression databases (Gene Expression Omnibus (GEO): www.ncbi.nlm.nih.gov/geo) show that it is transcribed in many human tissues (Table 4). We confirmed the latter observation by detecting OMP mRNA in two tonsil biopsies and in a lymphocytary pellet (Suppl Fig. 1A). Therefore, we considered as olfactory epithelium positive (OE+) the samples that did not lie within the 99% confidence limits derived from the three inferior turbinate mucosa biopsies (i.e., 798 cDNA copies/50 ng RNA, corresponding to the black line in Fig. 1b). On this basis, 11 samples of our series (five septa and six turbinates) were considered as OE+, while 21 samples (11 septa and 10 turbinates) were considered as olfactory epithelium negative (OE−).

Table 3 OMP and TAAR gene expression in nasal mucosa biopsies
Full size table
Table 4 GEO search results on hOMP expression
Full size table

Transcripts of hTAAR1, hTAAR2, hTAAR6, and hTAAR9 were detected in a few samples (respectively nine samples for hTAAR1, three samples for hTAAR2, three samples for hTAAR6, and five samples for hTAAR9) but always at trace levels, i.e., below the linearity range of the assay (less than 10 cDNA copies/50 ng RNA). hTAAR8 was detected in five samples at trace levels and in another five samples at higher levels, ranging from 11 to 33 cDNA copies/50 ng RNA. The most relevant results concerned hTAAR5, which was expressed in 11 samples at levels ranging from 15 to1,480 cDNA copies/50 ng cDNA (mean ± SEM was 293 ± 126). See Fig. 1c for a graphic summary.

hTAAR1 was expressed both in OE+ and in OE− samples and even in two samples obtained from nonolfactory mucosa. All the other hTAARs were expressed only in the OE+ group, and the difference was statistically significant by contingency analysis (P < 0.05 for hTAAR2, P < 0.01 for hTAAR5, P < 0.05 for hTAAR6, P < 0.01 for hTAAR8, and P < 0.01 for hTAAR9). Notably, TAAR5 was expressed in all OE+ samples, although no significant correlation was observed between TAAR5 expression and OMP expression (r = 0.393, P = NS).

The sampling site did not appear to affect the results. The percentage of septal and turbinate biopsies providing OE+ samples were similar (5/16 vs 6/16, P = NS by contingency analysis). No significant difference was also observed with regard to the frequency of TAAR-positive samples in the different locations (Suppl Table 1).

Despite the tissue heterogeneity of OE biopsies and the inter-individual variability of OE distribution, we tested whether TAAR expression was related to clinical variables like age, sex, and medical history, but no significant correlation was observed.

Discussion

This is the first report that TAAR genes are expressed in the human olfactory epithelium. In the past investigations on humans, patients have been limited by the difficulty in sampling olfactory areas within the nose. With the advent of endoscopy, biopsy sampling techniques have become safe and free from major complications. The new technique proposed by Lanza et al. (1993) improved the success rate of the biopsy and eliminated adverse effects, as confirmed in our experience.

A critical issue is the exact identification of the human olfactory neuroepithelium, whose anatomic localization is variable and poorly defined, since the borders of the olfactory organ are interspersed with islands of respiratory tissue. Inter-individual variability is particularly large because of the continuous loss of olfactory neurons due to inflammatory, infectious, or chemical injury, which is followed by regenerative processes leading to olfactory epithelium replacement by patches of respiratory epithelium or to olfactory epithelium invasion into regions of respiratory epithelium (Escada et al. 2009).

For these reasons, we could not rely on conventional histology to identify the olfactory epithelium. Preliminary investigations performed in two patients confirmed that histological detection of olfactory neuronal cells in a slice from nasal mucosa biopsy cannot assure that the adjacent epithelium contains olfactory neurons (data not shown).

Since OMP protein is specifically expressed in olfactory neurons (Weiler and Benali 2005), we planned to consider the presence of the corresponding mRNA as an olfactory epithelial marker. However, OMP mRNA was detected in all our samples, and unexpectedly, even in the three biopsies taken from the inferior turbinate mucosa, where the presence of the olfactory epithelium can be excluded (Escada et al. 2009). We interpret these results by speculating that the OMP gene may be transcribed at a low level in nonneuronal cells of the nasal mucosa, in spite of the absence of detectable amounts of OMP protein. This hypothesis is supported by several lines of evidence:

  1. 1)

    Analysis of the Gene Expression Omnibus database (www.ncbi.nlm.nih.gov/geo) showed that OMP mRNA was expressed in several normal human tissues, including epithelia (small and large airways, kidney, skin, and esophagus) and lymphocytes (Table 4). This was confirmed by our detection of OMP mRNA in two tonsil biopsies and in a lymphocytary pellet.

  2. 2)

    Dissociation between mRNA expression and protein expression can be accounted for by post-transcriptional regulatory mechanisms. For instance, in rat olfactory mucosa, the mRNAs coding for several connexins have been detected in the absence of the corresponding proteins, which has been attributed to the effect of multiple microinterfering RNAs (Weiler and Eysel 2003).

For these reasons, we decided to classify as OE+ only those biopsies in which OMP expression exceeded the upper 99% confidence limit of the values obtained in the inferior turbinate mucosa samples. While this threshold may be arbitrary, it should be pointed out that our results would be virtually unchanged by using different criteria, e.g., 95% confidence limits or the 90th percentile.

TAAR1 has been originally identified as a novel G protein-coupled receptor able to interact with trace amines like para-tyramine or β-phenylethylamine but not with the classical biogenic amines, namely catecholamines, histamine, or serotonin (Borowsky et al. 2001; Bunzow et al. 2001; Grandy 2007). Other closely related receptors were later identified and collectively called TAARs since most of them are not actually activated by trace amines when expressed in heterologous cell models (Zucchi et al. 2006; Lindemann et al. 2005; Lewin 2008). While another possible endogenous TAAR ligand is represented by 3-iodothyronamine (Scanlan et al. 2004), it has been recently reported that most mTAARs can be activated by several volatile amines and are specifically located in the olfactory epithelium, suggesting that they may play an important role in olfaction (Liberles and Buck 2006; Liberles 2009).

The present investigation is the first to be performed in human subjects. Our data suggested that all TAARs except TAAR1 are located in the olfactory epithelium and are absent in nonolfactory nasal mucosa. Liberles and Buck also maintained that expression of all mTAAR subtypes except for mTAAR1 is a specific feature of the olfactory mucosa, if compared to the other mouse organs. While this statement has been disputed (see review by Zucchi et al. 2006), we could not access different human tissues, so we can only confirm that hTAAR expression, and particularly hTAAR5 expression, is specific for olfactory vs nonolfactory nasal mucosa. Protein expression studies would obviously add relevant information, but specific ligands or effective antibodies for TAAR subtypes different from the rat TAAR1 are not available yet (Xie et al. 2007; Bunzow et al. 2001)

In mouse, Liberles and Buck used quantitative PCR and reported expression levels ranging from 500 to 5,000 copies/50 ng RNA. The expression of hTAARs was much lower, since only hTAAR5 and hTAAR8 exceeded trace levels, and the highest expression was in the order of 300 copies/50 ng RNA for hTAAR5. The difference may be due to the much greater thickness and continuity of the mouse olfactory epithelium, which is associated to a much higher sensitivity of the olfactory system (Escada et al. 2009).

It should be pointed out that while mouse TAARs are circumscribed in specific zones of OE, this feature has not been demonstrated in human OE. Therefore we cannot exclude that different TAARs may be preferentially expressed outside the current tissue collection zone.

mTAARs have been reported to respond to several volatile amines, and it has been speculated that they may have a specific role in the regulation of behavior, since there is evidence that mTAAR3, mTAAR4, and mTAAR5 may be involved in detecting pheromones or other urinary social cues (Liberles and Buck 2006; Liberles 2009). In mouse, TAAR5 can be activated by male mouse urine, with a half maximal response at 30,000-fold urine dilution. mTAAR5 activators include trimethylamine (EC50 = 0.3 μM) and dimethylethylamine (EC50 = 0.7 μM). The former is enriched in male compared to female urine (about 10–30 fold) and reaches a concentration of 5–10 μM in adult mice. This would be consistent with the putative role of the main olfactory and vomeronasal systems in mating behavior and fertility (Hagino-Yamagishi 2008).

Interestingly, hTAAR5 appears to be the major subtype in human subjects, while hTAAR3 and hTAAR4 cannot be expressed since they are pseudogenes. Trimethylamine is considered a maleodorant compound, since it is a product of plants and animal decomposition, and it is responsible for the characteristic smell of decaying fish (Mitchell and Smith 2001). It is liberated by gut microbes from precursors within the diet and is excreted by breath, urine, feces, sweat, and other bodily secretions (Mitchell and Smith 2001). Trimethylamine excretion is remarkably increased in primary trimethylaminuria, also known as “fish-odor-syndrome” (Mitchell and Smith 2001), a genetic disease due to complete or partial inactivation of the enzyme flavin monooxygenase 3 (FMO3), responsible for the N-oxidation of trimethylamine into the odorless and nonvolatile N-oxide. Trimethylamine production also depends on sex, and FMO3 appears to be regulated by sexual hormones in women (Shimizu et al. 2007; Hukkanen et al. 2005). Evidence of menstrual cycle-dependent variations of the olfactory threshold has also been reported (Navarrete-Palacios et al. 2003). Although this is only circumstantial evidence, the potential role of trimethylamine as a sort of pheromone deserves further investigation.

Obviously, the potential role of hTAARs in the regulation of human behavior is limited by the fact that man is a microsmatic mammalian, in which behavioral responses are largely based on other senses rather than on smell. In addition, steroid compounds derived from sexual hormones, which are found in human axillary sweat and urine, probably prevail over amines in producing pheromone-like effects (Savic et al. 2001).

Conclusions

We provided for the first time evidence that different TAAR subtypes, particularly hTAAR5 and hTAAR8, are specifically expressed within the olfactory area of the human nasal mucosa. These receptors might be involved in olfaction and particularly in the detection of volatile amines. Further studies are necessary to clarify the functional role of this receptor system and the underlying transduction pathways.