7 human breast cancer tissues were obtained from Prof. Dr. Tannapfel, Institute for Pathology, Ruhr-University Bochum. The study was carried out in accordance with the Declaration of Helsinki. A written informed consent was signed by each patient, who participated. The experiments were approved by the Regional Research Ethics Committee (Ethical Committee of the Medical Faculty, Ruhr-University Bochum); Ethic no: 17-6113.

Details on breast carcinoma tissues:

Tumor tissues were first preserved in RNAlater (Qiagen, Hilden, Germany) and then disruption was carried out using a Homogenizer (Precellys24, Bertin, Montigny-le-Bretonneux, France) and 1.4 ceramics beads (Precellys^®). The RNeasy Mini Kit (Qiagen, Hilden, Germany) was used according to the manufacturer’s protocol to isolate the RNA from the tumor samples. In addition, gDNA-Eliminator columns (Qiagen, Hilden, Germany) and an additional on-column DNaseI digestion were used according to the manufacturer’s protocol to avoid genomic contaminations.

The expression of the ORs detected via RNA-Seq was validated using RT-PCR. RNA was reverse transcribed using the iScript cDNA Synthesis Kit (Bio-Rad Laboratories, Hercules, CA, USA) according to the manufacturer’s instructions. An equivalent of 50 ng of RNA was used for each RT-PCR experiment. The primers detected 200–300 bp of the OR ORF. GoTaq qPCR Master Mix (Promega, Madison, WI, USA) was used according to the manufacturer’s protocol for the PCR reaction. PCR was performed using the Mastercycler realplex² (Eppendorf, Hamburg, Germany) (20 µl total volume, 40 cycles: 95°C, 59°C, 72°C, 45 s each). All experiments were conducted in triplicate.

Primer sequences:

Basic NGS analysis was performed as previously described (26). Briefly, for transcriptome analysis, RNA from tissue samples was prepared as stated above. At the Cologne Center for Genomics Next Generation Sequencing unit, libraries for NGS were constructed from mRNA. RNA-Seq was performed on the Illumina HiSeq2000 sequencing platform as paired-end reads with 101-nucleotide length. Raw data were generated for seven breast carcinoma tissue samples in the fastq format. Reads were aligned to the human reference genome hg19 by using Tophat (v1.2.0) (27). Bowtie, a short-read mapping program, was used to arrange the alignments and includes TopHat (28). SAMtools was used to sort and index the output files in the BAM format (29). The calculation of the abundance of the aligned reads was conducted by the program Cufflinks (30). Cufflinks uses the RefSeq reference transcriptome in the Gene Transfer Format obtained from the University of California Santa Cruz Genome Bioinformatics database. The relative abundance of the transcripts for each gene was quantified by Cufflinks using the unit FPKM (fragments per kilobase of exon per million fragments mapped) (27). The visualization of the data sets was conducted using the Integrative Genomic Viewer (http://software.broadinstitute.org/software/igv) to analyze the read distribution.

Data sets used for the RNA-Seq analysis of various cell lines, tumor tissues and healthy tissues were obtained from the SRA archive. Data sets of brain (31–33), data sets of breast (34–39), data sets of lung: SAMN00194128, data sets of pancreas (40), Accession numbers PRJNA237340 and PRJNA30709.

Hana3A cells were used for the recombinant expression of OR2B6 and were kindly provided by H. Matsunami (Duke University Medical Center, Durham, NC, USA). Hana3A cells were cultured in DMEM containing 10% fetal bovine serum, 5 mM Glutamax and 100 units/ml penicillin/streptomycin (Gibco^®, Life Technologies, Carlsbad, CA, USA). The cells were cultured in a 5% CO₂ humidified atmosphere at 37°C and were passaged every 3–4 days at a confluence of 90% using TrypLE^TM (Life Technologies, Carlsbad, CA, USA).

For the expression of ORs in Hana3A cells, expression vectors were constructed using standard PCR methods. ORs were expressed as fusion proteins containing 20 amino acids of the N-terminal region of the bovine rhodopsin (rhodopsin tag) as described elsewhere (41).

To ensure the specificity of the OR antibodies, we used Hana3A cells to express recombinant ORs. Cells were grown on cover slips in 24-well plates and were transfected with Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. The transfection mixture contained 300 ng of the OR plasmid and 60 ng of the RTP1s plasmid per plate, which enhances the transport of the OR into the cell membrane.

The specificity of hOR2B6 antibody (obtained from Abcam, dilution 1:50) was analyzed using immunocytochemistry of OR2B6-transfected Hana3A cells. The OR2B6-transfected cells were grown on coverslips in 24-well cell culture plates and fixed in 4% paraformaldehyde at 4°C for 20 min. Then, cells were washed and permeabilized in PBS^−/− + 0.1% Triton X-100, and blocking was conducted in PBST + 1% fish gelatin for 1 h at RT. The purified goat IgG polyclonal antibody against OR2B6 (Abcam, dilution: 1:100 in blocking solution) as well as the monoclonal anti-rhodopsin antibody 4D2 (anti-rho; Abcam; dilution 1:100) were incubated for 2 h at RT. After that, three washing steps were performed with washing solution, and the secondary antibodies coupled to Alexa Fluor 488 (Invitrogen) and 633 were applied for 45 min at RT. Cells were washed three times and covered with Prolong^® Gold Antifade reagent (Life Technologies). Images were taken with a Zeiss LSM 510 Meta confocal microscope. The localization of OR2B6 in Hana3A cells was visualized using a purified goat IgG polyclonal antibody against OR2B6 (Santa Cruz, dilution: 1:100 in blocking solution) and mouse monoclonal anti-goat antibody (Abcam, dilution 1:1000 in blocking solution). 4′6-diamidino-2-phenylindole (dilution 1:200) was applied for staining of the cell nuclei.

Breast cancer tissue samples were derived from patients undergoing radical mastectomy or punch biopsies. Samples were immediately stored at −80°C for subsequent RNA isolation or immunohistochemistry. Immunohistochemistry of human breast carcinoma tissue paraffin sections was performed on a Ventana BenchMark Ultra instrument (Ventana Medical System, Tucson, AZ, USA). For staining, the UltraView Universal DAB Detection Kit was used according to the manufacturer’s protocol. Primary antibody was diluted 1:50 and incubated for 32 min. Furthermore, all sections were co-stained with hematoxylin to illustrate tissue architecture. Staining was visualized by an Olympus BX 43 microscope.

All results were first tested for normality (Shapiro–Wilk test) and equal variance. If the data had a normal distribution, a two-tailed unpaired Student’s t-test was used. Data that were not normally distributed were analyzed using a Mann–Whitney test or a Kruskal–Wallis one-way analysis of variance. The statistical significance is indicated as follows: *p < 0.05, **p < 0.01, and ***p < 0.001.

The development of the RNA-Seq technique has revolutionized quantitative expression profiling and contributed substantially to the understanding of ectopically expressed ORs in diverse human tissues (15, 16, 20, 26, 42, 43). This advancing technology led to the accumulation of a large number of transcriptome data sets, which are available for reanalysis for a variety of mRNA-based questioning. We reanalyzed the RNA-Seq data of different human breast carcinoma tissues and breast carcinoma cell lines from the SRA archive to set up an expression profile of ORs in breast carcinoma tissues in comparison to corresponding normal tissues. Furthermore, we collected different breast carcinoma tissue samples and generated RNA-Seq data of seven breast cancer tissues samples. We used the Illumina RNA-Seq protocol to generate 5–10 million reads per transcriptome, with a read length of 101 bp. We analyzed all data using TopHat and Cufflinks software as described in the study by Flegel et al. and mapped the sequence reads to the human reference genome (hg19) (26). Expression intensities of all detected transcripts were calculated using the FPKM value (44). Transcript structures and read distribution were analyzed by visualizing the read alignment using the IGV. We analyzed the expression of ORs in 66 breast carcinoma tissues and cell lines and seven healthy breast tissues and ranked them according to their expression values (Figure 1). One receptor, OR2B6, had a pronounced expression in over 80% of all breast carcinoma tissues and 70% of all breast carcinoma cell lines, whereas it is not expressed in normal breast tissues (Figure 1C). OR2B6 was considered as “expressed” when FPKM value was >0.1. Expression values of OR2B6 ranged from FPKM 0.1 to 10.6, whereas an FPKM of 10.6 is remarkably high for an OR in non-olfactory tissues. OR2A7 and OR2W3 showed the second and third highest expression in the ranking, but there is, in contrast to OR2B6, a moderate expression in the healthy tissues as well. Beside OR2B6, OR56A3, OR4K15, OR2T8, OR4N4, OR8G5, OR2B2, OR13H1, OR7D2, and OR52N4 showed no expression in normal breast tissues, but were not uniformly expressed or even not expressed at all in most breast carcinoma tissues and cell lines.

We detected OR2B6 in human breast carcinoma cell lines and in breast carcinoma tissues, but not in healthy tissues (Figure 1).

To quantify the expression of OR2B6, we analyzed the expression of OR2B6 in a larger amount of healthy breast tissue data (n = 214, origin: GTEx database) (Figure 2C). We confirmed the expression of OR2B6 in breast carcinoma tissues, whereas it was hardly detectable in healthy breast tissues (Figure 2C). Then, we analyzed the expression of OR2B6 in other healthy human cell lines (Figure 2A) and healthy tissue samples that originated from the BodyMapProject (Figure 2B) (26). Here, OR2B6 showed transcript expression with FPKM value <0.5 in 4 of 16 tissue samples. In addition, we investigated the expression in 52 different healthy tissues, originating from the GTEx database (Figure 2D). OR2B6 was only detectable in Epstein–Barr virus-transfected lymphocytes (median RPKM, 1.5), prostate (median RPKM, 0.3), spleen (median RPKM, 0.3), testis (median RPKM, 0.4), and whole blood (median RPKM, 0.3). Notably, 111 ORs are detectable in testis and several in Epstein–Barr virus-transfected lymphocytes (26).

Thereafter, we focused on the expression of OR2B6 in breast carcinoma cell lines and cell lines that originated from other types of carcinoma (Figure 3). Here, 43 of 45 breast carcinoma cell lines do express OR2B6 transcripts (Figure 3A). OR2B6 is detectable in cell lines originated from other carcinoma types as well (Figure 3B), such as colon and gastric carcinoma cell lines, with FPKM values of up to 0.86, in a pluripotent embryonal carcinoma cell line (FPKM value: 2.16) as well as in a melanoma cell line (FPKM value: 1.42). In lung carcinoma tissue samples, there is a weak to moderate expression in two samples.

To validate the NGS sequencing data of OR2B6 transcript expression, we confirmed the expression of OR2B6 in native breast carcinoma tissues via RT-PCR (Figure 4A) and observed the distribution of mapped reads using IGV. Thereby, we detected an interesting read distribution. Both the coding sequence of OR2B6 and a histone transcript, HIST1H2BO, were covered with reads and are connected by a splice junction (Figure 4B). HIST1H2BO is a member of the histone cluster 1 H2B family. Family member E of the same cluster was found to be relevant in endocrine-resistant breast cancer (45). ORs are capable of building fusion transcripts with transcripts from upstream genes (26). We carried out an RT-PCR with exon spanning primers between OR2B6 and HIST1H2BO and could verify the expression of this fusion transcript (Figure 4C). Interestingly, this fusion transcript was detected in all breast carcinoma tissue samples via RT-PCR, whereas we could not detect it in healthy tissues although there was one exception. Our analyses show the expression of two transcripts. First, the expression of a transcript coding for OR2B6, which was supported by reads covering the complete coding sequence and the detection of the transcript with PCR primers P3 and P4. Second, the expression of a HIST1H2BO splice variant, which contains parts of OR2B6 as a downstream exon, detected by primers P1 and P2 and by exon spanning reads in the NGS sequencing. We analyzed the OR2B6 protein in native breast carcinoma tissue samples by performing immunohistochemical staining of breast carcinoma tissue samples using a specific antibody detecting OR2B6. Specificity of OR2B6 was confirmed by immunocytochemistry of OR2B6-transfected Hana3A cells using antibodies detecting OR2B6 and a rhodopsin tag located in the OR2B6-containing plasmid (Supplement Figure S1 in Supplementary Material). Consistently, OR2B6 expression was detectable in breast carcinoma cells (Figures 4D–F).

Thus, our data indicate that OR2B6 expression is associated with breast carcinoma and therefore might be considered as a potential tumor marker. Further analyses need to elucidate whether OR2B6 has a functional role in breast carcinoma as well.

While detecting OR2B6 transcripts in several cell lines that originated from different carcinoma tissues (Figure 3), we investigated the expression profile of OR2B6 in lung, pancreas, and brain carcinoma tissue samples. For the analysis of the OR expression profile in lung carcinoma, 5 normal lung tissues and 20 lung carcinoma tissues were investigated. In total, 20 ORs were ranked due their expression level (Figure 5A). Strikingly, the ORs expressed in the healthy lung in general show a higher expression level than the ORs in lung carcinoma tissues. OR2A4/7 shows, on average, the highest expression in most of the cancerous samples, with FPKM values ranging from 0.2 to 5.0. OR51E1 is detectable in both normal and cancerous tissues, whereas some ORs, such as OR1F12 and OR10AD1, are predominantly expressed in healthy lung tissues. Interestingly, OR6A2 and OR10A5 are only expressed in healthy tissues, with one exception.

Analysis of the expression pattern of ORs in 20 pancreas carcinoma tissues and cell lines revealed increased expression of OR4C6 (Figure 5B). This expression was detectable in 12 of 20 samples, with an FPKM value of 0.1–6.2. The verification of OR4C6 expression via the IGV is shown in Figure S2 in Supplementary Material. By analyzing the expression of OR4C6 in 52 healthy human tissues from the GTEx portal, it was striking that OR4C6 showed a median expression of 0.0 RPKM in all tissue samples (data not shown), which indicates that it may represent a possible biomarker for pancreatic carcinoma. The second highest expression was observed for OR2B6, which was expressed in 10 of 20 pancreatic carcinoma samples. Interestingly, the expression of both receptors was restricted to cancerous tissues. OR7E24 and OR2A7 were also expressed, but were not specific to carcinoma tissues in comparison to normal pancreas tissues.

Furthermore, we observed the expression of ORs in brain tissue and brain tumor tissue (Figure 5C). It was striking that there were only three ORs (OR4F21, OR1F1, and OR2B6) expressed in cancerous tissues and cell lines, whereas several ORs were expressed in three different healthy brain tissues (Figure 5C). OR14I1 and OR2L13 showed the highest expression, with FPKM values ranging from 1 to 10 (Figure 5C).