Specific and reliable detection of Myosin 1C isoform A by RTqPCR in prostate cancer cells

Cell lines

One normal prostate epithelial cell line (RWPE-1), prostate cancer cell lines LNCaP, 22Rv1 and PC3, A431 epidermoid carcinoma, A549 lung carcinoma, and non-cancer 3T3 fibroblasts were obtained from American Type Culture Collection (ATCC, http://www.atcc.org, Manassas, VA, USA). RWPE-1 cells were cultured in DMEM (Paneco, Russia), with 4.5 g/L glucose, containing 10% FBS (Gibco, Carlsbad, CA, USA), 1% gentamycin and supplemented with 0.05 mg/ml bovine pituitary extract (BPE) and 5 ng/ml human recombinant epidermal growth factor (EGF). Cancer cell lines and 3T3 fibroblasts were cultured in DMEM/F12 (Paneco, Russia) supplemented with 10% of FBS, 1% gentamycin and L-glutamine. Cells were grown in a humidified atmosphere with 5% CO₂ at 37 °C.

RNA extraction and cDNA synthesis

Total RNA extraction was performed with RNeasy Mini Kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s instructions, RNA samples were treated with DNAse I (Fermentas). RNA concentrations were measured on Nanophotometer (Implen, Germany), and RNA purity was confirmed by A260/280 and 260/230 ratios. 1 µg of total RNA was taken into the reverse transcription reaction with iScript Advanced cDNA synthesis kit (Bio-Rad, Hercules, CA, USA) according to manufacturer’s protocol.

RT qPCR

Real-time qPCR experiments were performed on CFX96 Touch cycler (Bio-Rad, Hercules, CA, USA). All samples were processed in triplicate. One sample of cDNA put into each PCR run served as an inter-run calibrator for uniting data into one experiment. Primer details are provided in Table S1. All amplicon sequences included at least one exon-exon junction to avoid DNA amplification. Primers were purchased from Synthol (Russia). Primer specificity was confirmed by melting curve analysis and detection of products with predicted length on 1.5% agarose gel electrophoresis. Amplification efficiency E was calculated as E = [10(−1∕slope) − 1], using the slope of the semi-log regression plot of Cq versus log input of cDNA. Each reaction was performed in triplicate. The reaction and cycling conditions were performed as described elsewhere (Potashnikova, Gladkikh & Vorobjev, 2015).

The Ct values were determined for real-time PCR curves by setting the threshold at 5 SD for each run. Relative cDNA quantity was calculated as ${\displaystyle \mathbf{Q}={\mathbf{E}}^{c\left(q\right)min-C\left(q\right)n},}$ where E—PCR efficiency, C(q)n—averaged triplicate C(q) value for each sample, C(q)min—minimal average C(q) value for the gene in the experimental set.

Results were analyzed using geNorm (Center of Medical Genetics, Ghent University hospital, geNorm version 3.5, 2002) and NormFinder (Molecular Diagnostic Laboratory, Department of Clinical Biochemistry, Aarhus University Hospital, Denmark, 2004).

Immunofluorescence

Mouse monoclonal antibody against N-terminal peptide of myosin 1C isoform A was described in Ihnatovych et al. (2012) (Ihnatovych et al., 2012). For myosin 1C isoform A visualization, cells were fixed with 4% paraformaldehyde in PBS, pH 7.2, for 15 min at 4 °C, washed three times with PBS, permeabilized with 0.1% Triton X100 and 0.1% Tween 20 in PBS for 1 h at room temperature, stained with first antibodies (1:150 dilution), then with second anti-mouse antibodies conjugated with Cy2 (Sigma-Aldrich, St. Louis, MO, USA) (dilution 1:100) and DAPI (Sigma-Aldrich, St. Louis, MO, USA) for nucleus visualization. Specimens were mounted in Mowiol 488 (Sigma-Aldrich, St. Louis, MO, USA). Fixed cells were imaged on Nikon TiE fluorescent microscope under PlanApo 60 ×/1.4 objective (phase contrast). Images were recorded by Hamamatsu ORCA FLASH2 digital camera, using FITC (Ex. 450–490 nm; Em. 510–540 nm) and DAPI (Ex. 340–380 nm; Em. 435–485 nm) filter sets. Preparation conditions and exposure time was kept constant for all specimens used.

Image analysis

Data were analyzed using ImageJ software (NIH, Bethesda, MD, USA). Quantitative measurements were performed on the unprocessed original 16-bit B/W images. To perform the analysis of mean fluorescence intensity values, we picked the small image regions of the same area inside the nucleus, in the cytoplasm and outside the cell (these values were taken for the signal-to-noise ratio), calculated the average fluorescence intensity values and performed signal to noise correction. Fluorescent images were processed using ImageJ and finalized with Adobe Photoshop (Adobe Systems, San Jose, CA, USA) software.

Flow cytometry

PC3 and RWPE-1 cells (at approx. 50% confluence) were detached using trypsin then consecutively washed with 10% FBS-medium and PBS. Detached cells were stained for flow cytometry in PBS according to antibody manufacturers’ protocols (15 min at +4 °C, in the dark). Antibody details were as follows: anti-CD44-BV421 (clone BJ10, BioLegend), anti-CD90-BV510 (clone 5E10, BioLegend), anti-CD133/1-PE (clone AC133, Miltenyi Biotec), anti-CD57-PerCp-Cy5.5 (clone HNK-1, BioLegend), anti-CD24-APC-H7 (clone ML5, BD). Cells were washed with PBS and analyzed using a 6-laser FACSAria SORP instrument (BD Biosciences, USA).

Median fluorescence intensity (MFI) values were obtained from FACSDiva 6.1.3 software for each surface marker under study in the overall PC3 and RWPE-1 cell populations. The overall cell populations were gated based on light scatter properties and median signal intensities were exported for them. Immunophenotypes of the model cell lines were monitored several times throughout cell passaging. Average MFIs and their SDs were calculated for surface markers from different passage data.

Data analysis

Statistical data and graphs were obtained using GraphPad^® Prism (GraphPad Software, version 5; San Diego, CA, USA). Data were presented as medians and ranges or as median ± SEM. Differences between data sets were tested with the Mann–Whitney U test, inter-group differences with p < 0.05 were reported as significant.

Immunophenotype of prostate cell lines

First, PC3 and RWPE-1 model cell lines were immunophenotyped by flow cytometry. Both cell lines are characterized as CD44+/CD90-/CD133-/CD57-/CD24+- when subcultured under passage 8 (Fig. 1).

However, some markers (CD44 and CD24) showed high MFI (Mean fluorescence intensities) variance between passages in both PC3 and RWPE-1 cell lines. This can be best illustrated by the temporal heterogeneity of CD44 expression during RWPE-1 early passages (Fig. S1): here we show an increase in a stem cell marker CD44 expression upon passage 4. A minor CD133-positive subpopulation appears at passage 8, while CD90 remains absent at all passages.

To access the MFI variance, average MFI values for each marker and their SDs were calculated for both cell lines based on three immunophenotypes from different passages. For PC3 cells they were as follows: CD44 MFI = 9489.67 ± 8605.98, CD90 MFI = 172 ± 43.35, CD133 MFI = 101.67 ± 34.59, CD57 MFI = 78.67 ± 12.58, CD24 MFI = 136.50 ± 57.28. The average MFI values for RWPE-1 cells were as follows: CD44 MFI = 9567.67 ± 7871.92, CD90 MFI = 176.33 ± 95.03, CD133 MFI = 137.67 ± 37.63, CD57 MFI = 121.67 ± 32.32, CD24 MFI = 142.00 ± 50.91.

Both cancer and non-cancer prostate cell lines exhibited the same set of surface markers, thus flow cytometry as a sole method was not enough to distinguish prostate cancer cells. For better evaluation of the differences between cancer and non-cancer cells we employed RT qPCR.

Optimal reference genes for prostate cell lines

The choice of external reference controls for normalization and qPCR data analysis is critical in case of low mRNA quantities. Reference genes have to be expressed ubiquitously in normal and malignant tissue, and their expression should not be affected by experimental conditions. The reliable set of reference genes has to include high abundance genes (that correspond to transcriptional levels of MYO1C A isoform in cancer cells) and low abundance genes (that correspond to transcriptional levels of MYO1C A isoform in normal prostate). A sufficient number of genes for qPCR data normalization also have to be estimated.

Initial screening confirmed that all candidate genes were expressed in prostate cell lines. Raw Cq values of the chosen genes indicate that the list of candidates includes genes with high (HPRT1), moderate (ACTB, TBP, YWHAZ, RPL13A, B2M, ALAS1) and low (UBC, SDHA, GAPDH) abundance (Fig. S2). qPCR analysis was verified by the presence of a single sharp peak on the melting curve and a single product band of predicted size on agarose gel electrophoresis. For all candidate genes, efficiency values ranged between 0.93 and 1.05 and R² values ranged between 0.992 and 0.999 (Table S2). We also evaluated the expression of the candidate reference genes for the cells on different passages and showed that mRNA expression of candidate genes in RWPE-1 and PC-3 lines did not change at least for 10 passages, as standard deviation of the normalized cDNA quantity for all genes did not exceed 0.04 (Table S3). Expression stability of the selected genes was determined with GeNorm and NormFinder algorithms.

In GeNorm, we ranked the selected genes according to their expression stability (M value). All candidate genes except for B2M and RPL13A passed the 1.5 stability threshold value proposed by GeNorm. GeNorm ranking order indicated that YWHAZ/GAPDH pair had the lowest M value (0.52). HPRT1 and ACTB genes also had stability values below 1 (Table 1). Unlike GeNorm, NormFinder does not evaluate paired variances and estimates the independent stability values (SV). NormFinder ranking established GAPDH as the most stable single gene across all cell lines under study (SV = 0.08) and HPRT1 as the second stable gene with SV = 0.11, while YWHAZ had SV = 0.13 (Table 1).

To estimate the optimal number of genes for qPCR data normalization we employed pairwise variation analysis proposed by GeNorm. Pairwise variations are defined as SD of log2-transformed expression ratios of any pair of reference genes required for geometric mean normalization. As shown in Fig. 2, the lowest V values were for V4/V5 for all cell lines except LNCaP, V4/V5 was also the lowest for the total set, indicating that addition of the fourth gene to normalization factor (NF) is unnecessary. Put together, our data imply that a combination of YWHAZ, GAPDH, and HPRT1 serves best for mRNA expression analysis in the set of prostate cell lines.

mRNA levels of myosin 1C isoform A in malignant and benign prostate cell lines

To prove the prostate cancer-associated expression of myosin 1C isoform A we analyzed its expression in prostate cancer cell lines (PC3, LNCaP, and 22Rv1) and compared it to non-cancer prostate cell line RWPE-1. We confirmed the previous data of Ihnatovych, Sielski & Hofmann (2014) and showed that relative normalized cDNA quantity of myosin 1C isoform A in prostate cancer cells is at least two orders of magnitude higher than in non-cancer prostate cells. Medians of relative normalized cDNA quantity in prostate cancer cell lines varied from 1.938 (PC3) to 3.426 (22Rv1) compared to 0.037 in RWPE-1 cells. Myosin 1C isoform A was not expressed in normal fibroblasts (3T3 cell line, median normalized cDNA quantity 0.003) and demonstrated low expression levels in non-prostate cancers, like A431 epidermoid carcinoma (median mRNA expression 0.022) or A549 human lung carcinoma (median mRNA expression 0.011) (Fig. 3, Table 2). Myosin 1C A isoform expression for PC3 and RWPE-1 did not change significantly for different passages; relative normalised quantity of myosin 1C isoform a on the 10th passge was 2.012 for PC3 and 0.045 for RWPE-1, respectively.

To evaluate the detection limit of myosin 1C isoform A with real-time qPCR, we performed a series of dilutions of PC3 cells by RWPE-1 cells as follows: 1:1, 1:3, 1:10, 1:30, 1:100, 1:300 and 1:1,000 (where one part corresponded to 10⁴ PC3 cells). Normalized relative cDNA quantity of myosin 1C isoform A varied from 1.478 (1:1 mixture) to 0.078 (1:1,000 mixture) compared to 2.389 for pure PC3 cells and 0.021 for pure RWPE-1 cells (Fig. 4). Thus, qPCR evaluation of myosin 1C isoform A expression can be considered a sensitive and selective method for prostate cancer detection, even for low quantities of cancer cells in the specimen.

Protein expression of myosin 1C isoform A is prostate cancer-specific

To confirm that myosin 1C isoform A is a specific marker of prostate cancer at protein level, we performed immunofluorescence staining of PC3 and RWPE-1 cells using specific antibodies against myosin 1C isoform A. In PC3 cells we observed the bright staining of myosin 1C isoform A throughout the cytoplasm and in the nucleus (Fig. 5A), whereas in RWPE-1 cells the staining was much weaker, although the protein was also located both in the nucleus and in the cytoplasm (Fig. 5B). Further staining of A549 cells with the antibody against myosin 1C isoform A also revealed a weak signal both in the nucleus and in the cytoplasm (Fig. S3).

For quantitative evaluation of protein expression, we performed mean fluorescence intensity calculations for myosin 1C isoform A staining in the nucleus and cytoplasm on 58 PC3 cells and 56 RWPE-1 cells. Median MFI of myosin 1C isoform A staining for nuclei of PC3 cells was 1749.0 compared to 924.9 in RWPE-1 cells (differences were statistically significant, p < 0.0001). The differences in cytoplasmic staining of myosin 1C isoform A were less distinct, median MFI of the cytoplasmic signal was 813.5 for PC3 cells and 602.1 for RWPE-1 cells (differences were statistically significant, p < 0.0001) (Fig. S4).