Chromogenic in situ hybridization reveals specific expression pattern of long non-coding RNA DRAIC in formalin-fixed paraffin-embedded specimen

Long non-coding RNA (lncRNA) plays an important role in the regulation of gene expression in normal and cancer cells. We previously discovered a novel tumor-suppressive lncRNA, DRAIC, in prostate cancer cells. Subsequent studies have demonstrated that DRAIC is dysregulated in various malignancies and exhibits a tumor-suppressive or pro-oncogenic function. However, details regarding its expression pattern in normal and cancerous tissues remain largely unknown. In this study, we performed chromogenic in situ hybridization (CISH) using RNAscope technology to assess DRAIC expression in formalin-fixed paraffin-embedded (FFPE) specimens. In the neuroendocrine-differentiated cancer cell line VMRC-LCD, CISH revealed a diffuse localization of DRAIC in the cytoplasm as well as specific accumulation in the nuclear compartment. DRAIC expression was comprehensively analyzed using tissue microarrays containing 89 normal and 155 tumor tissue samples. DRAIC was weakly expressed in normal epithelial cells of the colon, bronchiole, kidney, prostate, and testis. Conversely, DRAIC was moderately to highly expressed in some cancer tissues, including prostate adenocarcinoma, invasive ductal carcinoma of the breast, neuroendocrine carcinoma of the esophagus, lung adenocarcinoma, and small cell lung carcinoma. While DRAIC knockdown did not affect VMRC-LCD cellular viability and invasive ability, gene expression related to the neuroendocrine and cancer-related pathways was altered. Our expression analysis revealed the specific expression pattern of DRAIC in normal and cancerous FFPE tissues. The results presented here may lead to the elucidation of additional novel functions of DRAIC.


Introduction
Recent advances in transcriptomic technology have revealed that the vast majority of the human genome is transcribed to produce non-coding RNA, which is not translated into proteins [1].Along with well-characterized conventional non-coding RNAs such as tRNA and rRNA, various other types of non-coding RNAs are intracellularly expressed.Among these, long non-coding RNAs (lncRNAs) are defined as RNAs longer than 200 nucleotides in length with no open reading frame.While the total number of lncRNA genes remains unclear, it has been estimated that the number exceeds 100,000 in a human cell [2].Although a few lncRNAs have been functionally characterized, some have been demonstrated to play biological roles in physiological processes, including development, cellular differentiation, and homeostasis [3].Therefore, the abnormal expression of a specific lncRNA is closely linked to the initiation and progression of diseases such as cancer.
Each lncRNA is expressed in the nucleus, cytoplasm, or both and can interact with other molecules, such as nucleic acids and proteins, to perform its functions.For example, nuclear lncRNAs (e.g., NEAT1 and MALAT1) are involved in chromatin remodeling, gene transcription, and RNA splicing through their interaction with transcription factors and RNA-binding proteins [4,5].Cytoplasmic lncRNAs (e.g., TINCR and LINC00152) regulate the mRNA and protein turnover, modulate molecular signaling activity, and sequester microRNAs (miRNAs) [6,7].
Because the functions of lncRNAs are dependent on their subcellular localization and binding partners, it is essential to identify their distribution inside a cell to determine their biological roles.
Previously, we identified a novel tumor-suppressive lncRNA, DRAIC (Downregulated RNA in cancer, inhibition of cell invasion and migration), which is a 1.7 kb spliced and polyadenylated RNA that was predominantly expressed in the cytoplasm of prostate cancer cell lines [8].We demonstrated that DRAIC had tumor-suppressive effects and its high expression was a good prognostic indicator in patients with prostate adenocarcinoma [8].Subsequently, DRAIC was revealed to interact with subunits of the IκB kinase (IKK) complex to inhibit NF-κB activity and suppress prostate cancer progression [9].Furthermore, the tumor-suppressive mechanisms of DRAIC have been reported in glioblastoma [10], gastric adenocarcinoma [11], retinoblastoma [12], and triple-negative/basal subtype of breast carcinoma [13].Conversely, pro-oncogenic functions of DRAIC have been demonstrated in nasopharyngeal carcinoma [14] and lung adenocarcinoma [15].Although the molecular mechanisms of the contradictory roles of DRAIC have not been fully elucidated, they might be attributed to its versatility in regulating multiple pathways that include autophagy [10,16], protein translation [10], ubiquitin modification [11], and immune cell infiltration [17].To understand its complex mode of action [18], it is important to analyze the expression level and subcellular localization of DRAIC in each cell line and tissue.
Archived formalin-fixed paraffin-embedded (FFPE) samples are invaluable resources for identifying pathological biomarkers.Chromogenic in situ hybridization (CISH) is an analysis method to visually detect RNA in FFPE tissues at the single-cell level.We have demonstrated the efficacy of CISH to analyze the miRNA expression patterns in normal and cancerous FFPE tissue samples [19,20].Although various in situ hybridization methods for lncRNAs have been developed [21], detecting lncRNAs in FFPE specimens remains challenging because the longer transcript is easily degraded during formalin fixation [22].Furthermore, CISH technology with increased sensitivity is required to visualize lncRNAs because their expression level is typically lower than that of mRNAs [23].To overcome these limitations, we performed CISH utilizing RNAscope technology [24] to detect DRAIC in FFPE cell lines and tissue microarrays (TMAs).The RNAscope CISH successfully visualized DRAIC expression, thus providing novel insights into the role of DRAIC in a subset of malignancies.

Cell culture
The VMRC-LCD, 22Rv1, and PC3 cell lines were obtained from the Japanese Collection of Research Bioresources Cell Bank (Osaka, Japan), the European Collection of Authenticated Cell Cultures (Salisbury, UK), and the American Type Culture Collection (Manassas, VA, USA), respectively.VMRC-LCD cells were maintained in Dulbecco's Modified Eagle Medium (DMEM) (Thermo Fisher Scientific, Waltham, MA, USA) containing 10 % fetal calf serum (FCS) (Thermo Fisher Scientific), and penicillin/streptomycin (P/S) (FUJIFILM Wako Chemical Corporation, Osaka, Japan), while the 22Rv1 and PC3 cells were maintained in RPMI-1640 (Thermo Fisher Scientific) containing 10 % FCS and P/S.
For the pathological analysis, the cells were washed twice using phosphate-buffered saline (PBS) (Thermo Fisher Scientific) without trypsin treatment and fixed using 10 % neutral buffered formalin (FUJIFILM Wako Chemical Corporation) for 3 h at room temperature (RT).The fixed cells were embedded in paraffin after being gelled with 1 % sodium alginate and 1 M CaCl 2 , as previously reported [25].The 3-μm thick sections were mounted on silane-coated slides (New silane 3) (MUTO PURE CHEMICAL CO., L.T.D., Hongo, Tokyo, Japan).

Cell viability assay
VMRC-LCD cells at 1 × 10 4 cells/well were seeded in a 96-well plate.At 24 h after seeding, siRNA transfection was performed as described above.The cells were subjected to a cell viability assay at the designated time point using WST-8 (DOJINDO LABORATORIES, Kumamoto, Japan) according to the manufacturer's instructions.The 450-nm absorbance of samples was measured using Sunrise Rainbow (Tecan, Männedorf, Switzerland).

Matrigel invasion assay
VMRC-LCD cells at 1 × 10 5 cells/well after transfecting siRNA were seeded into 24-well Matrigel Invasion Chamber (#354480, Corning, NY, USA) in serum-free DMEM.Ten percent FCS was added only to the lower compartment.After incubation for 48 h, the noninvaded cells were removed from the upper surface of the membrane by scrubbing with a swab.The invaded cells were fixed and stained with 0.5 % crystal violet in 20 % methanol solution for 15 min, and counted per membrane.

Gene expression microarray and Gene Ontology (GO) analysis
Total RNA from VMRC-LCD cells transfected with si-NC and si-DRAIC2 was isolated using NucleoSpin RNA.The RNA integrity was assessed by an Agilent Technologies 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA), and gene expression analysis was performed using the Human Clariom S Assay (Affymetrix, Santa Clara, CA, USA).The raw data were obtained as CEL files, which were processed using the Signal Space Transformation Robust Multi-Chip Analysis (SST-RMA) algorithm in the Transcriptome Analysis Console software (Thermo Fisher Scientific) with background correction, quantile normalization, and summarization.The expression dataset was deposited into the National Center for Biotechnology Information Gene Expression Omnibus (NCBI GEO) (https://www.ncbi.nlm.nih.gov/geo/) and is accessible through GEO Series accession number GSE242691.GO analysis was performed using Metascape (https://metascape.org/gp/index.html#/main/step1)[26], with the upregulated (fold change >1.5) or downregulated (fold change < − 1.5) genes in si-DRAIC2 compared with si-NC.

Chorioallantoic membrane (CAM) assay
The CAM assay was performed in accordance with the protocol that was approved by the Animal Affairs Committee at Fujita Health University (approval number APU22067).The fertilized eggs were obtained from the Fukushima branch of MORI BREEDING FARM (Fukushima, Japan).The CAM assay was performed as previously reported, but with modifications [27].Briefly, the eggs were placed in the incubator (Rcom MX-20, BellBird, Japan) set at 37.5 • C, with 65 % humidity, and the eggs were rotated once an hour.On day 8 of incubation, egg rotation was stopped, and a round window was opened using the round bar to expose the CAM.PTEF O-ring (Sansho, Tokyo, Japan) was placed on the xenografted area of the CAM, and VMRC-LCD cells at 5.0 × 10 6 in a total of 20 μL medium (10 μL PBS and 10 μL Matrigel basement matrix (Corning)) were xenografted.The shell window was covered with Tegaderm™ (3 M Japan Limited, Tokyo, Japan), and the egg was placed back into the incubator.On day 5 of the CAM assay, a tumor on the CAM was excised and the eggs were sacrificed in a − 20 • C freezer for 2 h.The tumor was fixed using 10 % neutral buffered formalin for pathological analysis.

CISH
CISH was performed on the cell lines and TMA samples (FDA331, FDA808I-1, FDA808I-2, EN801b, and LC811a) (TissueArray.Com LLC, Derwood, MD, USA) with RNAscope (R) 2.5 HD Reagent Kit-BROWN (Advanced Cell Diagnostics (ACD)) (Newark, CA, USA).The Z probes for DRAIC (#568281) and DapB (negative control, #310043) were obtained from ACD.The Z probe for DRAIC that targeted 2-1270 of NR_026979.1 was designed and synthesized by ACD.The commonly available TMAs analyzed in this study have been widely used in research, and their academic value has already been determined.Therefore, the Institutional Review Board deemed that this study was not required for ethical review in light of the Ethical Guidelines for Medical and Biological Research Involving Human Subjects.
After deparaffinization, the samples were treated with H 2 O 2 to block endogenous peroxidase activity.The target retrieval was performed for 3 min using the SR-MP300-K (Panasonic, Kadoma, Osaka, Japan).The samples were treated with the Protease Plus reagent for 15 min (cell lines) or 30 min (TMAs).Each Z probe was hybridized for 2 h at 40 • C in a HybEZ Hybridization System (ACD).The signal was amplified using AMP1 (30 min at 40 • C), AMP2 (15 min at 40 • C), AMP3 (30 min at 40 • C), AMP4 (15 min at 40 • C), AMP5 (60 min at RT), and AMP6 (15 min at RT) reagents.The signal was visualized using the 3,3′-diaminobenzidine (DAB) chromogen diluted in the DAB substrate buffer, and the slides were counterstained with hematoxylin.
Individually stained slides were scanned with Axioscan7 (Oberkochen, Germany) and visualized using QuPath open-source software [28].DRAIC expression was analyzed by counting dark brown dots regardless of the dot size in both the nucleus and cytoplasm of 100 cells from 3 random fields of view and was indicated as the average [29].Instances where the specimen detached during the CISH procedure were excluded from analysis.Instances with a large amount of hemosiderin and other biological pigments were also excluded from analysis because it was challenging to distinguish them from the DRAIC signal.Each slide was evaluated by two independent board-certified pathologists (K.S. and S⋅Y.).

Expression of DRAIC in FFPE cell line block samples and xenograft tissues
We performed CISH using RNAscope, a technique to detect intracellular RNA with high sensitivity and specificity [24].After two paired Z probes were tandemly hybridized to the target RNA, the L-shaped preamplifier binds to the upper region of the Z probes (Fig. 1a).The amplifiers then bind to multiple sites on each preamplifier, and the labeled probes containing a chromogenic enzyme bind to multiple sites on each amplifier, which makes it possible to detect low levels of RNA with high sensitivity.A pool of multiple paired Z probes can detect partially fragmented RNAs in an FFPE specimen.When the paired Z probes do not tandemly bind to RNA, the preamplifier does not bind, thus ensuring high specificity [24].
We first evaluated the expression of DRAIC in human cancer cell lines.Based on the DRAIC expression analyzed using the Cancer Cell Line Encyclopedia (CCLE) dataset [30], we selected three human cancer cell lines: VMRC-LCD, 22Rv1, and PC3 cells (Fig. S1).RT-qPCR showed that VMRC-LCD had the highest expression of DRAIC, 22Rv1 had moderate expression, and PC3 had the lowest expression (Fig. 1b).We performed the RNAscope CISH for DRAIC using the FFPE samples, which revealed that the strongest signal and largest number of dots were found in VMRC-LCD (Fig. 1c, left), weaker and fewer signals in 22Rv1 (Fig. 1c, center), and the weakest and least amount of signals in PC3 (Fig. 1c, right) (Fig. S2).As for the subcellular localization, diffuse small dots were predominantly observed in the cytoplasm, while some accumulated larger dots were detected in a nuclear compartment.As a negative control, we confirmed that no signal of DapB, a bacterial RNA, was observed in these three cell lines (Fig. S3).
We next detected DRAIC in the VMRC-LCD xenograft in the CAM assay.After xenografting VMRC-LCD cells onto the chick CAM, a macroscopically visible tumor formed at day 5 (Fig. 1d).CISH detected DRAIC signals in the VMRC-LCD xenograft but not in the chick CAM tissue (Fig. 1e).Similar subcellular localization patterns shown in Fig. 1c were observed, namely many diffuse small dots in the cytoplasm and larger dots in the nucleus.

Expression of DRAIC in TMAs
Based on the results using the FFPE cell block samples and xenograft tissues, we performed the RNAscope CISH on the TMAs to comprehensively evaluate the expression pattern of DRAIC in normal and tumor tissues.In Fig. 2 and Supplementary Tables, the results of normal tissues, various tumor tissues excluding lung cancer, and lung cancer tissues were summarized.In normal tissues (n = 89), DRAIC was weakly expressed in some epithelial cells of the colon, bronchiole, kidney, prostate, and testis and was mostly localized in the nucleus (Figs.2a and  3, and Table S1).DRAIC was not expressed in non-epithelial cells such as fibroblasts and blood cells.

Potential downstream genes of DRAIC in the neuroendocrine carcinoma-derived cell line, VMRC-LCD
Our recent findings based on in silico analysis revealed that a higher expression of DRAIC is associated with the neuroendocrine (NE) feature in lung cancer [31].Because CISH detected the DRAIC signal in a subset of NECa and SCLC (NE-differentiated lung carcinoma) (Figs. 2 and 4), we hypothesized that DRAIC may perform a biological role in NECa.Thus, we transfected siRNAs against DRAIC in VMRC-LCD, which is a NE-differentiated LUAD cell line [32].Although two different siRNAs (si-DRAIC1 and si-DRAIC2) against DRAIC reduced DRAIC expression compared with the negative control (si-NC) (Fig. 5a), the cellular viability and invasive ability were not affected by DRAIC knockdown (Fig. 5b and Fig. S5).To investigate the potential downstream genes, we selected si-DRAIC2 for gene expression microarray because it showed the slightly greater effect on DRAIC knockdown than si-DRAIC1 (Fig. 5a).We found 1100 upregulated and 1343 downregulated genes in si-DRAIC2 compared with si-NC-treated VMRC-LCD cells (Fig. 5c).GO analysis showed the enrichment of GO terms such as brain development, regulation of nervous system development, cell maturation, MAPK, and cell migration in the downregulated genes (Fig. 5d).On the other hand, GO terms including regulation of hormone levels and regulation of cell-substrate adhesion were enriched in the upregulated genes (Fig. S6).

Discussion
Studies on the expression of DRAIC have been mainly conducted by RT-qPCR using total RNA from cell lines and clinical samples.Publicly available transcriptome data such as microarray and RNA-sequencing have also been used to analyze DRAIC expression [8,31].Although these approaches are beneficial in understanding its overall expression tendency, detailed information regarding its expression patterns (e.g., cell type specificity) is not available.Furthermore, even in the worldwide databases, such as The Cancer Genome Atlas, not all benign and malignant histological types are collected.
In this study, we utilized RNAscope technology to detect DRAIC expression in FFPE specimens.RNAscope is an established technology to quantitatively analyze the expression of coding and non-coding RNA with high sensitivity and specificity [24,29,33,34].Using this technology, we successfully detected DRAIC in FFPE cell block samples and xenograft tissues, and these expression levels were consistent with the CCLE dataset and our RT-qPCR results.VMRC-LCD, which had the highest expression levels of DRAIC in the analyzed cell lines, had many small dots in the cytoplasm.Since we showed in our cell fractionation assay that DRAIC was predominantly expressed in the cytoplasm of prostate cancer cells [8], several molecular mechanisms as a cytoplasmic lncRNA have been reported [9,10,14,15,35].Importantly, this study revealed that DRAIC was also expressed in the nuclear compartment as larger dots.Although the precise region where the nuclear DRAIC was expressed has not been identified, this finding may lead to the understanding of novel functions of DRAIC, considering that nuclear lncRNAs regulate gene expression through multiple mechanisms [4,5].Tiessen et al. demonstrated through cell fractionation analysis that DRAIC was expressed both in the nucleus and cytoplasm of a breast cancer cell line, MCF-7 [16].These results indicate that the pattern of DRAIC subcellular localization might be cell-type dependent.While the molecular mechanisms of this unique pattern remain unclear, the sequence motif and its binding partners might affect the localization [36].It would be important to analyze the intracellular distribution in each cell line by RNAscope CISH to determine the diversity of biological functions.This is the first report analyzing the DRAIC expression on TMAs.In normal tissues, some epithelial cells weakly expressed DRAIC, mainly in their nucleus.In contrast, some specific cancer cells had moderate to high DRAIC expression in their nucleus and cytoplasm, although the distribution pattern varied.In addition to the adenocarcinomas of the prostate and breast, whose DRAIC expressions have been analyzed [8,13,37], this study demonstrated the high expression in a subset of NECa, LUAD, and SCLC.Our previous bioinformatics-based analysis indicated that DRAIC was highly expressed in ASCL1 (a NE-differentiation-associated transcription factor)-positive LUAD and SCLC [31].Because RNAscope could be simultaneously performed with immunohistochemistry within the same specimen [38], it would be interesting to further demonstrate the positive correlation between DRAIC RNA and ASCL1 protein in these instances.In addition, by analyzing the infiltrating cell types, we plan to clarify whether DRAIC is an immune-related lncRNA because it was suggested to be involved in the immune cell infiltration in LUAD [17].It would also be important to analyze DRAIC involvement in the prognosis of patients with LUAD because contradictory results have been presented in our [8,31] and other [15] reports.
We performed a gene expression microarray after DRAIC knockdown by siRNA in VMRC-LCD, an NE-differentiated LUAD cell line.The GO terms related to NE were enriched by knockdown, which suggests that DRAIC might be involved in the characteristics of cancer cells with NEdifferentiation, although the molecular mechanisms are unknown.DRAIC knockdown also influenced the gene expression associated with cellular migration and cell-substrate adhesion, while cell viability was not affected.Consistent with these results, DRAIC knockout using the CRISPR/Cas9 system induced cell migration and anchorageindependent growth in prostate cancer cells, while cell proliferation at the standard plate was not altered [9].Notably, DRAIC knockdown remained low invasive ability in VMRC-LCD cells, suggesting that the effect of DRAIC on cellular invasion is also cell type-dependent.Because DRAIC induces or represses cellular invasion, depending on cancer cell types [8,9,15,39], we should analyze the cellular phenotypes in vitro and in vivo using various NE-differentiated cancer cells.In addition, given that siRNA strategies are not effective in knocking down nuclear RNAs because RNAi-machinery is localized in the cytoplasm [40,41], we need

Conclusion
We successfully detected DRAIC expression by RNAScope CISH in FFPE cell line samples and TMAs.Because this study aimed to establish RNAscope CISH for DRAIC and investigate expression patterns in various cancers, the case numbers of each histological type were limited.Whether and how DRAIC expression is changed during carcinogenesis and involved in the prognosis of patients will be investigated next using a larger number of FFPE specimens.Although further analyses are required to fully understand DRAIC expression and function, our results provide novel insights into the molecular mechanisms of DRAIC in physiological and disease conditions.

Informed consent statement
Tissue microarrays commercially obtained from TissueArray.Com LLC were developed in compliance with their ethical standards with the donor's informed consent and privacy.

Fig. 1 .
Fig. 1.Expression of DRAIC in formalin-fixed paraffin-embedded (FFPE) cell line block and xenograft samples, (a) Schematic illustration of chromogenic in situ hybridization (CISH) using RNAscope.The RNAscope steps are shown in brief on the right.(b) The expression of DRAIC in VMRC-LCD, 22Rv1, and PC3 cell lines was measured by RT-qPCR and normalized to HPRT1.Mean ± standard deviation (S.D.) n = 3.The expression in VMRC-LCD was set as 1. (c) DRAIC staining of VMRC-LCD (left), 22Rv1 (center), and PC3 (right) was analyzed by CISH.Scale bar: 50 μm.(d) Chorioallantoic membrane (CAM) assay.Upper panel: the window was opened and a PTEF O-ring was placed on the CAM.Lower panel: the tumor mass (white arrow) of the VMRC-LCD xenograft on the chick CAM at day 5. Scale bar: 10 mm.(e) DRAIC staining in the CAM assay was analyzed by CISH.The region with a black line rectangle at the left panel is magnified right.Scale bar, 100 μm (left) and 50 μm (right).

Fig. 2 .
Fig. 2. Expression of DRAIC in tissue microarrays (TMAs), (a)-(c) The expression of DRAIC was analyzed by CISH in normal tissues (a), various tumor tissues (excluding lung cancers) (b), and lung cancer tissues (c).Each bar indicates the average number of dots per 100 cells from three random fields of view.DLBCL; diffuse large B-cell lymphoma, NECa; neuroendocrine carcinoma, GIST; gastrointestinal stromal tumor, panNET; pancreatic neuroendocrine tumor.

Fig. 3 .
Fig. 3. Representative images of DRAIC expression in normal tissues, Representative images of DRAIC expression in normal epithelial cells of the colon (a), bronchiole (b), prostate (c), and testis (d) tissues.Arrows indicate some of the DRAIC signals.Scale bar: 50 μm.The region with a black line rectangle is magnified below.The length of each magnified panel is 50 μm.

Fig. 4 .
Fig. 4. Representative images of DRAIC expression in cancerous tissues, Representative images of DRAIC expression in NECa of the esophagus (a), lung adenocarcinoma (LUAD) (b), and small cell lung carcinoma (SCLC) (c).Scale bar: 50 μm.The region with a black line rectangle is magnified below.The length of each magnified panel is 50 μm.

Fig. 5 .
Fig. 5. Effect of siRNA-mediated knockdown of DRAIC in a neuroendocrine-differentiated lung cancer cell line, VMRC-LCD, (a) The expression of DRAIC after transfection with siRNAs against DRAIC in VMRC-LCD cells was measured by RT-qPCR and normalized to HPRT1.Mean ± S.D. n = 3.The expression of the sinegative control (NC) was set as 1. *p < 0.05.(b) The cell viability of si-NC, si-DRAIC1, and si-DRAIC2 of the VMRC-LCD cells was measured using a WST-8 assay.The viability at 0 h of each siRNA-transfected cell was set as 1. (c) Heat map showing the upregulated (FC > 1.5) and downregulated (FC < − 1.5) genes of si-DRAIC2 compared with si-NC in the VMRC-LCD cells.FC; fold change.(d) Gene Ontology terms for the Biological Processes enriched in the downregulated genes of si-DRAIC2.