RNA recovery from specimens of duct-washing cytology performed contemporaneously with mammary ductoscopy

Conventional cytological diagnosis including duct-washing cytology (DWC) is sometimes performed using ductal epithelial cells collected during mammary ductoscopy; it is useful for detection of early-stage breast cancer such as ductal carcinoma in situ (DCIS). However, conventional cytological diagnosis focuses exclusively on cellular morphology; false negatives and false positives may be caused by inadequate specimen preparation (triggering cell degeneration) or poor examiner diagnostic skills. Molecular diagnosis using RNA biomarkers is expected to compensate for the weaknesses of cytological diagnosis. We previously employed microarray analysis to identify highly expressed genes in DCIS, suggesting that they may be useful for DCIS diagnosis. Here, we explored whether DWC samples yielded RNA of sufficient quantity and quality for RNA biomarker-based diagnosis. We extracted RNAs from 37 DWC samples. RNA from 12 samples exhibited RNA integrities of  ≥ 6, indicative of moderate-to-high quality. We then showed that cocaine and amphetamine regulated transcript prepropeptide (CARTPT) and breast cancer-associated transcript 54 (BRCAT54) mRNA—previously shown by microarray analysis to be highly expressed in DCIS—were detectable in these samples. Therefore, DWC samples may be useful for molecular diagnosis involving RNA biomarkers.


Introduction
Breast cancer (BC) originates in the mammary ducts. Some cancers may be metastatic, thus progressing from non-invasive to invasive. Because of substantial advances in breast imaging tools, ductal carcinoma in situ (DCIS) constitutes approximately 20% of newly diagnosed BCs; its proportion is increasing [1,2]. Early BC detection improves prognosis. Abnormal nipple discharge (ND) may be the earliest presenting symptom of BC; substances in such discharge may yield detailed information regarding BC. Indeed, carcinoembryonic antigen in ND has served as a BC biomarker [3,4]. At present, ND is only regarded as an aid to diagnosis; cytological and/or histological evaluation is considered essential for accurate diagnosis. Preoperative cytological diagnosis can be accurate if atypical cells are present in small samples. Ultrasound-guided fine needle aspiration biopsy is widely used to collect material from the ventral surfaces of mammary glands [5]. However, conventional cytological diagnosis focuses exclusively on cellular morphology; false negatives and false positives may be caused by inadequate specimen preparation (triggering cell degeneration) or poor examiner diagnostic skills.
Mammary ductoscopy (MD) allows direct visualisation of the ductal lumen using a fibreoptic microendoscope, as well as the collection of cytological specimens for ductwashing cytology (DWC); this is an alternative form of preoperative BC diagnosis. It would be clinically useful if surplus DWC specimens could be used for RNA extraction and subsequent molecular diagnosis involving RNA biomarkers. Here, we explored whether RNA of sufficient quantity and quality could be recovered from DWC specimens of DCIS patients. Furthermore, we explored whether the RNA contained cocaine-and amphetamineregulated transcript prepropeptide (CARTPT) and breast cancer-associated transcript 54 (BRCAT54) mRNAs, which a previous microarray study found to be highly expressed in DCIS [6].

Surplus specimens of duct-washing cytology
The surplus tissues of 45 specimens collected for DWC at Musashi Kosugi Hospital from December 2017 to October 2019 were used. All patients presented with ND. Observation and specimen sampling were carried out as follows. First, nipple duct dilatation was performed using bougies; an 18-G Surflo intravenous catheter was then inserted. The duct was washed with a lidocaine solution, and the solution was collected. Second, Solid Fibrescope MS-611 (FiberTech Co. Ltd. Chiba, Japan) was inserted and the lesion was directly observed. MD allows direct observation of the ductal interior; a fibre of outer diameter 0.8 mm is placed within an 18-G Surflaw needle. Finally, aspiration biopsy was performed using an 18-G Surflo intravenous catheter after removal of the fibrescope [7,8].

Extract RNA from duct-washing cytology specimens
Total RNA was extracted from specimens using RNAiso-Plus (Takara-Bio, Kusatsu, Japan), generally in accordance with the manufacturer's instructions; however, all steps were performed below 4 °C and linear acrylamide (Thermo Fisher Scientific, Waltham, MA, USA) was added as a co-precipitant at the step of isopropyl alcohol precipitation (Additional file 1: Fig. S1). Each RNA sample was resuspended in 10 μL RNase/DNase-free distilled water.

Assessment of the quantity and quality of DWC specimens
RNA concentrations were determined using a Quanti-Fluor RNA System (Promega, Madison, WI, USA). The A260/A280 ratios of RNAs extracted from DWC specimens were determined using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific).

RNA quality assessment
The extracted RNA was further analyzed using the Agilent 2100 Bioanalyzer (Agilent, Santa Clara, USA) in combination with the RNA 6000 Nano Lab Chip kit (Agilent). The quality of RNA was assessed by RNA integrity number (RIN) [9].

Real-time reverse-transcription PCR
The levels of mRNAs encoding CARTPT, BRCAT54, and hypoxanthine phosphoribosyltransferase 1 (HPRT-1) (internal standard) were measured via quantitative reverse-transcription (qRT)-PCR using the SYBR Green Assay (Thermo Fisher). PCR products were monitored on a DICE Real-Time PCR System (Takara-Bio).

Immunobiological staining
Of the 37 patients who underwent MD, eight underwent surgery and seven turned out to have malignant tumors. For immunopathological diagnosis, they underwent immunobiological staining for estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) (Additional file 1: Table S1).

Statistical analysis
Spearman's rank correlation analyses between RNA yields and RIN were performed using EZR, a modified version of R software (version 1.53; R Foundation for Statistical Computing, Vienna, Austria) with additional biostatistical functions [10]. Figure 1 shows a flowchart of the selection of 37 patients. We obtained DWC specimens by MD and extracted RNA. Of the 37 patients, eight had malignancies or suspected malignancies; one patient's results were indeterminate. Of nine patients (eight with malignancies and one with indeterminate findings), two were diagnosed with DCIS, five were diagnosed with solid papillary carcinomas (SPCs), one was diagnosed with atypical ductal hyperplasia (ADH), and one had indeterminate findings; all underwent surgery, except the patient with indeterminate findings. One patient with suspected malignancy did not undergo surgery because she was under observation at another hospital. SPC is a particular subtype of DCIS that exhibits a full papillary growth pattern; it is described as a separate category of "intraductal papillary lesion" in the 4th edition of the World Health Organisation criteria [11]. ADH is a pre-cancerous state: a lesion that has both benign and malignant features. Intraductal papilloma (IDP), wart-like lump that develops in one or more mammary ducts, is regarded as a benign condition.

Results and discussion
Patient information including ages, DWC diagnoses, MD findings, and the RINs and amounts of isolated RNAs are shown in Table 1. Of the nine patients mentioned above, eight underwent surgery, and seven had malignancies (Additional file 1: Table S1). Immunobiological staining revealed that all seven were ER-and PRpositive with positive cell proportions of at least 80%; six malignancies were HER2-negative. The invasion grades were very low (0 or 0.3); the spread ranged from 2.0 to 10.5 cm (mean, 5.3 cm).
Of these, eight were malignant (six SPC and two DCIS); four were benign (four IDP). During the followup period, the tumours of two patients became malignant. Patient no. 16 was recently diagnosed with BC by puncture of a peripheral mass. Patient no. 27 eventually underwent surgery and diagnosed as mucinous carcinoma located in the peripheral region of the duct. We selected five samples (three malignant and two benign) with high RNA yields ( Table 2) for quantitative RT-PCR analysis of CARTPT and BRCAT54. These mRNAs were  chosen because microarray analysis previously showed that they were highly upregulated in DCIS-and SPCderived tissues [12,13] In addition, the levels of mRNA encoding HPRT-1 (a representative housekeeping gene) were analysed. RNA from the breast tumour tissue of an SPC patient-previously used in array analysis [6]served as the positive control (sample no. 0). CARTPT and HPRT-1 mRNAs were detected in all samples. BRCAT54 mRNA was detected in the samples although expressed at a quite low level in one IDP-derived sample (No.16) ( Table 2; Additional file 1: Fig. S2). Dissociation curves with single peaks were obtained for the samples (Additional file 1: Fig. S2), indicating that non-specific PCR amplicons were not generated.
To the best of our knowledge, no report about RNA extraction from MD samples has been published so far, although there are a few reports describing the analysis of DNA methylation in MD samples [14,15]. This study has clinical relevance in that we demonstrated a possibility that RNA of acceptable quality and quantity could be extracted from DWC specimens and applied to RNAbased pre-operative diagnosis of breast carcinoma.
In summary, we used MD to examine mass lesions in the mammary ducts and extracted RNA from DWC samples; many extracts exhibited moderate-to-high RNA quality. We found that mRNAs abundantly expressed in SPC or IDP can be detected via qRT-PCR.