T5224 br The study protocol was approved by the Institutiona
The study protocol was approved by the Institutional Review Board and the Hungarian National Ethical Review Committee ETT-TUKEB (2372/2012/EKU). All patients gave written informed consent for sample collection and epigenetic analysis before the onset of endoscopy. Sample and data management was done anonymously.
Sample procurement. Brush cytology samples were collected during ERCP procedures using a through-the-scope disposable gastrointestinal cytology brush (Ref# 152, ConMed, Utica, NY, USA). After preparation and fixation of direct smears for routine cytology analysis during the procedure, the distal brush tip containing the
samples was immediately positioned in 1.5ml empty Eppendorf tubes and stored temporarily in an IsoFreeze-Rack (KR-20B, Kisker Biotech GmbH & Co., Steinfurt, Germany) keeping the specimens at -10ºC to -20ºC, and were stored long-term in a -80ºC freezer.
Cytology analysis. The cytology specimens were first evaluated as part of the routine clinical work-up and were later reviewed independently by three pancreatobiliary cytology specialists who were blinded to the clinical records and the results of the genetic tests. The slides were scored on loss of honeycombing, T5224 clumping, increased nuclear/cytoplasmic ratio, nuclear molding, nucleoli, necrosis and inflammation . The smears were graded as negative (A), atypical (B), suspicious (C) and positive (D) for malignancy (Figure 1) . Throughout the study we consider only category D (high number of malignant cells) as a positive cytology result.
Total RNA isolation. Total RNA including the small RNA fraction was isolated, using a miRNeasy Mini Kit (Qiagen, Hilden, Germany) following the protocol provided by the manufacturer. The final volume of elution was 50µL. Measurements of RNA quality and quantity were performed with Qubit 1.0 fluorometer and Nanodrop (ThermoFisher Scientific, Waltham, MA, USA). Each sample was loaded on RNA Nano gel electrophoresis chips and visualized on the Agilent 2100 Bioanalyser (Agilent Technologies, Santa Clara, CA, USA) for quality of the small RNA band (from 25-200nt).
MicroRNA quantification by RT-PCR. Quantitative miRNA expression analyses were performed using the Taqman microRNA PCR Assay (ThermoFisher Scientific, Waltham, MA, USA). Reverse transcription of 8ng total RNA (per sample) to cDNA was converted according to the protocol of the Taqman microRNA Reverse Transcription Kit (ThermoFisher Scientific, Waltham, MA, USA). Quantitative RT-PCR was performed for hsa-miR-16-5p, hsa-miR-21-5p, hsa-miR-196a and hsa-miR-221-3p. Selection of analyzed microRNAs was based on existing literature suggesting high expression in pancreatobiliary cancer [14-18]. Small nuclear RNA U6 (RNU6B; Applied Biosystems, Foster City, CA, USA) was used as internal control. Primer microRNA sequences for the RT-PCR assay are provided in Supplementary Table 1. All real-time reactions were run on 96-well plates in duplicates, using the LightCycler 480 thermocycler (Life Science Roche). Statistical analysis All Ct (threshold cycle) values were normalized by subtracting RNU6B control values
(∆Ct), and depicted after subtraction from 40. Normalized expression values of 5 samples constantly remained out of
range for all microRNA markers studied, therefore these samples were excluded and further statistical calculations were performed on the final set of n=35 samples. Overexpression of specific microRNAs in malignant samples was
calculated using the ∆∆Ct method. We performed statistical analyses using the SPSS software (version 17.0, SPSS
Inc., Chicago, USA). ROC analyses were performed with the MedCalc software (version 15.6). Results with p<0.05 were regarded as statistically significant. Quantitative variables were described as mean ± standard deviations.
Malignant disease was present in 26 patients (74.3%), presenting as biliary and pancreatic strictures in 14 and 12 patients, respectively. Nine patients (25.7%) had benign pancreatobiliary disease causing stricture formation with prestenotic dilation and clinical symptoms necessitating an intervention. Routine cytology resulted in no false positive diagnoses (100% specificity), however, the true diagnosis of malignancy was missed in 46.2% of patients. If we considered suspicious (C) categories positive as well, the sensitivity increased somewhat (73.3%); however, this was accompanied by a decrease in specificity (not shown).
Expression of candidate microRNAs
The isolation of total RNA including microRNAs was successful in each of the analyzed brush cytology specimens, resulting in total RNA quantities of 2.4±1.8 µg per isolation on average with clean qualitative results showing one distinct peak (not depicted). Expression levels of all the analyzed target microRNAs were higher in the brush samples (n=35) of patients with malignant strictures compared to benign samples after normalization with RNU6B, however, this tendency did not reach statistical significance for miR-21 (p=0.1062) (Supplementary Figure 1.). No significant expression changes were observed comparing benign pancreatic and biliary samples, therefore we assigned them to a single normal group. Expression (∆Ct values) of miR-16 (p=0.0039), miR-196a (p=0.0003) and miR-221 (p=0.0049) showed a clear statistical significance between malignant and benign pancreatobiliary specimens (n=35), with median ∆Ct values differing (∆∆Ct) by 3.5 cycles for miR-16, 4.6 cycles for miR-196a, 2.1 cycles for miR-221, indicating approximately 11.3-fold, 23.6-fold and 4.3-fold increased microRNA levels in the isolates of malignant strictures compared to benign samples (Supplementary Figure 1.). Subsequent analyses are presented with the best three microRNA markers mainly (miR-16, miR-196a, miR-221). Baseline differences in gender distribution, pre-ERCP bilirubin levels, CA19-9 value and tumor size had no statistically significant effect on microRNA expression levels (not shown).