A Method to Pre-Screen Rat Mammary Gland Whole Tissue Mounts followed by RNAscope in situ hybridization

Background: RNA in situ hybridization is an extremely useful gene expression analysis technique that preserves the spatiotemporal nature of tissue and allows for evaluation of specic cell populations and morphologies. This technique is especially useful in evaluating expression changes in disease progression models. Tissue processing procedures used to identify pathological disease morphologies in situ could compromise RNA integrity and the reproducibility and quantitative accuracy of RNA in situ hybridization assays. Methods: A combinatorial approach to pre-screen rat mammary gland tissue whole mounts for hyperplastic and malignant lesions that were not visually discernible without staining was used. This pre-screening process was followed by an RNA in situ hybridization analysis method known as RNAscope. Results: We show that there are no differences in the quantitative nature of RNAscope assays between tissue that was immediately formalin-xed and paran embedded (FFPE), which is recommended by the manufacturer, and tissue that was whole mounted and pre-screened for lesions of interest prior to RNAscope. Conclusions: Preserving the integrity of RNA and quantitative nature of the RNAscope assay is important, as it allows unpalpable lesions to be directly identied, bypassing a need for labor-intensive serial sectioning of FFPE tissues to nd lesions. This method is applicable to any epithelial-based disease progression model using whole-tissue mounts. validate pairs either processed or whole mounted carmine prior to paran embedding. Tissue was probed for Mier3 family member 3 (Mier3), a moderately expressed gene, RNA Polymerase II Subunit A (Polr2a), a highly expressed gene, and bacterial 4-hydroxy-tetrahydrodipicolinate reductase (DapB), a negative control gene that should not be present in mammalian tissues. RNAscope probes per cell in non-diseased mammary ducts and mammary carcinomas were quantied using HALO-ISH Quantication software from Indica Labs. A two-way ANOVA followed by Tukey’s multiple comparisons test was performed using GraphPad Prism version 7.00 Software. In each tissue type, there was no signicant difference between processing methods (Panel A- p=0.5057, Panel B- p=0.6086). As expected, there was a signicant difference between probes in either non-diseased mammary tissue or mammary carcinomas (2-way ANOVA p<0.0001). signicantly different post hoc test (p<0.05).

Aluminum carmine is a biological nuclear stain derived from dried female cochineal insects [12]. When applied to mammary gland whole mounts, epithelial ductal structures are easily detected within mammary tissue cleared of fat [9]. This is useful for monitoring morphological changes during mammary carcinogenesis. It is relevant to visualize rat mammary gland ductal structures for pre-clinical research purposes because the vast majority of human breast malignancies arise from ductal epithelial cells [13][14][15]. Rat mammary carcinogenesis serves as a model of the natural histopathological lesions and malignancies found in human breast disease. Screening tissue whole mounts is a valuable technique to visualize morphological changes occurring during carcinogenesis [13,16].
RNAscope assays are designed and recommended to be used in recently processed formalin xed and para n embedded (FFPE) or cryopreserved tissue samples. There is a pervasive problem of nucleic acid degradation when handling tissue with any ISH assay, including RNAscope. Nucleic acid quality has effects on the reproducibility and quantitative analysis of downstream assays. Protein, DNA, and RNA are better preserved, for a longer period of time in FFPE tissues compared to frozen tissues [17,18].
Morphological features are also retained in FFPE tissues, which has made it a preservation method of choice for decades.
Immediate FFPE processing of tissue intended for RNAscope analysis is recommended to preserve the integrity of RNA. This is because the duration of xation and pre-treatment conditions can negatively impact nucleic acid integrity [19,20]; therefore, care must be taken to limit exposure time and harsh conditions prior to para n embedding. In this methods report, we outline how RNAscope can be effectively utilized following screening of mammary gland whole tissue mounts for lesions of interest. This is useful as it allows for identi cation and study of diseased tissue morphologies without apparent loss of RNA integrity. We show, using probes targeting genes of different expression levels, that the quantitative capacity of the RNAscope assay is retained following whole tissue mounting and processing. This is applicable to any epithelial-based disease progression model requiring pre-screening of whole tissue.

Protocol:
Animal husbandry Rats were maintained in an AAALAC-approved facility on a 12 h light/dark cycle and provided LabDiet 5001 Rodent Diet (PMI® Nutrition International) and water ad libitum. Inbred Wistar Furth (WF/Hsd) females were obtained from Envigo. All animal protocols were approved by the University of Louisville Animal Care and Use Committee.

DMBA Administration and Mammary Gland Tissue Resection
At 50-55 days of age 7,12-Dimethylbenz(a)anthracene (DMBA Acros Organics, CAS#: 57-97-6, 20 mg/mL in sesame oil) was given by single oral gavage (65 mg DMBA/kg body mass) to WF/Hsd (Envigo) females to induce mammary carcinogenesis. Mammary glands were resected four or twelve weeks following DMBA administration. Brie y, one skin incision was made at the lower abdomen and extended forward, along the midline, to the neck. Four additional skin incisions were made, each extending from the midline down each limb. Skin was separated from the body cavity to reveal mammary glands present in pairs along each ventro-lateral boundary of the body. The larger mammary glands, such as the combined thoracic pectoral glands toward the anterior (Fig 1, labeled B/C glands) or individual abdominal-inguinal glands toward the posterior (Fig 1, labeled D glands) were excised for whole mounting. Mammary gland tissue was carefully trimmed away from the skin, moving as far dorsally as possible. One abdominal D mammary gland was immediately processed by formalin xation and para n embedding (FFPE), while the other was whole mounted and processed as outlined below.

Whole Mount Preparation:
Immediately following tissue resection, mammary glands were stretched across an electrostatically charged slide and submerged in room temperature, 10% neutral buffered formalin overnight for xation. Slides were removed from formalin and placed into an acetone bath at room temperature overnight to defat. Overnight may not be enough time to entirely remove all fat; however, overnight is su cient and allows for sooner and more visible staining. Mammary glands will become translucent and may require up to 3 days in acetone to remove fat, depending on the thickness of the mammary gland.
Following defatting of the mammary glands, slides were submerged in aluminum carmine staining solution. Aluminum carmine solution was prepared by dissolving 1g of carmine alum and 2.5g of aluminum potassium sulfate in 500 mL of RNase-free water. Boiling 20 minutes was used to completely dissolve carmine into solution. Carmine solution was refrigerated for storage and reused after ltering to remove buildup of oil remaining in mammary glands. Staining may take 1-2 days depending on mammary gland thickness. Glands were considered fully stained when white tissue was no longer visible through the back of the slide. Whole mounted xed tissue not used for the purposes of RNAscope analysis can be stored long-term submerged in mineral oil.
Mammary gland tissue processing, embedding, and sectioning: To evaluate morphological structures and lesions within carmine stained mammary gland tissue whole mounts, such as hyperplasia and ductal carcinoma in situ (DCIS) (Fig 2), regions of interest were excised from whole-mounted glands using a surgical scalpel. Excised tissue was placed in histology cassettes and processed using a standard tissue processor. To properly embed tissue in para n blocks, tissues were oriented such that the at side, originally against the glass microscope slide, faced the bottom of the para n mold. Para n embedded tissues were stored at 4°C until ready to section. Messenger RNA is stable in para n blocks for effective use in RNAscope for up to 1 year [21].
Sectioning of FFPE blocks was done no more than one day prior to a planned RNAscope assay, as mRNA remains more stable within para n blocks than tissue sections [21]. Prior to sectioning, para n blocks were placed at -20°C for a minimum of 30 minutes to improve sectioning and reduce wrinkling of tissue sections. Tissues were cut into 4 mm sections and placed onto superfrost plus microscope slides. Other slide types, suitable for RNA in situ hybridization assays, may also be used. Slides were dried fully at room temperature overnight. Depending on tissue size, it was possible to t up to three sections per slide. This served to analyze multiple genes or controls on a single slide in an RNAscope assay.
Hematoxylin and Eosin (H&E) staining of sectioned mammary glands: An H&E stain on sectioned tissues was used to assess tissue sample quality and to con rm the lesion of interest was present in the section. Slides were depara nized and rehydrated using 3 separate xylene washes followed by single washes each in 100% ethanol, 90% ethanol, 80% ethanol, and DI water. Tissues were stained with hematoxylin for 3 minutes, rinsed under running DI water for 3 minutes, soaked in bluing reagent for less than a minute, and again rinsed with DI water. Slides were dipped a few times in 100% ethanol to eliminate water before staining in eosin-Y for 30-45 seconds. Tissue was then dehydrated using washes of 90% ethanol, 100% ethanol, and 3 separate xylene washes. Samples were air dried and coverslipped using permount or similar mounting media

RNAscope Assay
In this report, RNAscope assays were performed on pairs of D mammary glands, one immediately processed by FFPE and the other whole mounted, defatted, carmine stained, and processed and para n embedded later. The RNAscope 2.5 HD Detection Reagent-RED kit (Advanced Cell Diagnostics #322360) was used according to manufacturer's protocol except for modi cations to minimize harsh prehybridization treatment conditions on fragile mammary tissue. Slides were depara nized in two xylene, followed by two 100% ethanol washes. A hydrophobic barrier was drawn around tissue sections with an ImmEdge pen. Sections were blocked with H 2 O 2 for 10 minutes at room temperature. Slides were submerged in target retrieval solution at a temperature just below boiling for 10 minutes, rather than the manufacturer recommended boiling for 15 minutes. Protease plus solution was added to slides and incubated in HybEZ oven (ACD Bio) at 40°C for 10 minutes, rather than the manufacturer recommended 30 minutes. Target probe was added and allowed to hybridize for 2 hours at 40°C in the HybEZ oven.
Serial steps to amplify signal involved the following solutions and conditions: Amp1 for 30 minutes at 40°C, Amp2 for 15 minutes at 40°C, Amp3 for 30 minutes at 40°C, Amp4 for 15 minutes at 40°C, Amp5 for 30 minutes at room temp, Amp6 for 15 minutes at room temp, and detection reagent for 10 minutes at room temperature. Slides were washed in wash buffer between each probe hybridization and ampli cation step. Slides were counterstained with 1:1 Gill's hematoxylin and water, allowed to air dry, and coverslipped with mounting media. Slides were imaged on an Aperio ImageScope CS2.

Expression Quanti cation and Statistics:
Representative images were selected from n=6 mammary carcinoma and n=6 non-diseased mammary tissues that were either immediately FFPE processed or whole mounted prior to FFPE processing. Mean probe per cell was quanti ed for DapB, Mier3, and Polr2a expression using the HALO-ISH Quanti cation software from Indica Labs. Two independent DMBA groups were used to provide mammary gland or mammary carcinoma tissue, each tissue type was analyzed separately by Two-Way ANOVA. Independent variables to compare mammary gland and carcinoma expression levels were processing method and RNAscope probe. Probe per cell was the dependent variable. Tukey's post hoc test was used protect for multiple comparisons. Statistical analysis was performed using GraphPad Prism version 7.00 for Windows, GraphPad Software.

Results & Discussion:
Mammary gland whole mounts, are used to detect various stages of pathological mammary gland lesions, including pre-malignant hyperplasias, pre-invasive ductal carcinoma in situ (DCIS), and invasive ductal carcinoma (IDC). Standard histopathologic approaches, such as hematoxylin and eosin (H&E)staining of FFPE tissue sections, might miss morphological abnormalities not captured within evaluated tissue serial sections [11]. Whereas, pre-screening mammary gland whole-tissue mounts for abnormal morphologies allows for direct identi cation of lesions and hyperplastic regions. These regions can be excised and further processed separately for analysis by tissue sectioning.
Here, aluminum carmine stained whole tissue mount screening was used to detect mammary ductal epithelium morphological changes that occur during mammary cancer progression. Mammary ductal structures, interspersed within stromal tissue, are visible after staining with aluminum carmine [22]. Early hyperplastic proliferative lesions appeared, as expected, in aluminum carmine stained mammary gland whole mounts (Fig 2) [9,10,22]. These lesions were extended, thickened, and ampli ed ductal branches.
Ductal carcinoma in situ (DCIS), an intraductal proliferation of epithelial cells, (Fig 3)[13] appeared as darkened structures within whole mounts (Fig 2). This intense staining is due to a dense concentration of epithelial cells present in these lesions. The ability to detect and study epithelial-based diseased tissue morphologies within mammary glands following aluminum carmine staining is crucial since breast cancer predominantly arises from luminal epithelial cells lining ducts [13,14]. Invasive ductal carcinoma (IDC) presented in tissue sections as migratory epithelial cells penetrating ductal walls and invading stroma (Fig 3) [13,15]. These carcinomas appear as palpable mammary tumors that are visible without aluminum carmine staining.
Whole mount tissue processing, followed by RNAscope, was successfully used in this study to produce RNAscope results that were similar to results from samples directly processed by FFPE. Therefore, an exposure to additional reagents and time, which was required for whole tissue mounts processing prior to para n embedding, did not have a negative effect on mRNA quality or abundance. Four-hydroxytetrahydrodipicolinate reductase (DapB) is a bacterial gene that is not present in properly handled mammalian samples. No positive staining for DapB was observed using RNAscope in samples used in this study (Fig 3). This indicated no bacterial contamination occurred while processing whole mounts. To test the sensitivity of RNAscope following whole tissue mounting, we targeted the mesoderm induction early response family member 3 (Mier3) gene, a low to moderately expressed gene transcript in rat mammary and female breast tissue, with average expression of 11 transcripts per million [23]. For comparison, we probed for RNA Polymerase II Subunit A (Polr2A), a highly expressed gene with an average of 79 transcripts per million in human breast tissue [23]. RNAscope assays probing for Polr2a and Mier3 re ected these expected expression differences (Fig 3). Results observed for Polr2a were densely packed red puncta within mammary ducts, while a Mier3 probe resulted in more scattered red puncta. There was no discernable difference in puncta abundance between RNAscope assays performed on aluminum carmine stained mammary gland whole mounts and mammary glands that were formalin xed and para n embedded immediately following tissue resection (Fig 4). This lack of difference between tissue treatment methods, and an expected difference in transcript abundance between independent gene target probes, indicated that a pre-screening process did not negatively impact mRNA stability.
To verify that similar quantitative puncta density results were obtained from different tissue processing methods, RNAscope images were quanti ed using HALO-ISH software (Indica Labs) (Fig 5). Nondiseased mammary ducts displayed means ± standard deviation (SD) for DapB probe per mammary epithelial cell of 0.28 ± 0.26 by immediate FFPE and 0.35 ± 0.19 by tissue whole mounting. An RNAscope probe for Mier3 resulted in means ± SD per cell of 2.34 ± 0.46 in immediate FFPE and 2.32 ± 0.14 in tissue processed from whole-mounted mammary glands. The Polr2a probe yielded means ± SD per mammary epithelial cell of 7.55 ± 1.22 in immediately processed FFPE tissue and 7.21 ± 1.22 in whole mounted tissue. A two-way ANOVA indicated no statistically signi cant difference in quantities from tissues that were immediately FFPE processed according to the manufacturer's directions, or tissues that were whole mounted prior to FFPE processing. There was a statistically signi cant difference between gene probes, as expected (p<0.0001). This con rmed anticipated results of no to low DapB expression, moderate Mier3 expression in both tissue processing methods, and high Polr2a expression, which indicates the quantitative sensitivity of RNAscoppe assays was retained in whole mounted tissue. Similar results between tissue processing methods were obtained when RNAscope assays from mammary carcinomas were quanti ed. Means ± SD of the DapB probe per carcinoma cell were 0.15 ± 0.17 in immediate FFPE tumors and 0.23 ± 0.20 in whole mounted tissue containing a tumor. A Mier3 probe displayed means ± SD per cell of 4.45 ± 0.64 in immediate FFPE tumor tissue and 4.30 ± 0.59 in whole mounted tumor tissue. The Polr2a probe resulted in means ± SD per cell of 12.27 ± 1.65 in immediate FFPE processed carcinomas and 11.52 ± 1.90 in whole mounted carcinoma tissue. A two-way ANOVA was not statistically signi cant for tissue processing method, but was signi cant for assay probe (p<0.0001). These results validate that RNAscope is a robust and appropriate procedure for quantitative analysis of gene expression in whole mounted tissue followed by FFPE.

Conclusion:
RNAscope can be successfully used on FFPE tissue that was initially processed as whole mounted tissue. There was no loss in qualitative or quantitative values of RNAscope. This allows for pre-screening of potential lesions or morphological tissue changes that are only detectable with tissue staining. Crucially, methods outlined in this report provide a way to classify and separately process stages of disease progression while retaining the target speci city and robust quantitative nature of the RNAscope assay.  Premalignant and malignant lesions are visible in mammary gland tissue whole mounts following aluminum carmine staining. Shown are representative mammary glands from virgin female Wistar Furth (WF/Hsd) rats taken 8 weeks after administering a chemical carcinogen, 7,12-dimethylbenz(a)anthracene (DMBA). Hyperplasia, induced by DMBA, appears as dark thickened branching (demarked by black boxes). Carcinogen-induced DCIS appears as a dark mass within mammary fat pads (marked by black arrows).

Figure 3
Representative images of RNAscope using Mier3, Polr2a, and DapB probes on tissue sections obtained from rat mammary gland tissue whole mount preparations. Rat mammary gland ductal hyperplasia, ductal carcinoma in situ (DCIS) and invasive ductal carcinoma (IDC) were rst identi ed in formalin xed aluminum carmine stained whole-tissue mounts followed by excision and processing. The RNAscope assay performed on these morphologies used Mier3 family member 3 (Mier3) as a moderately expressed gene (left column), RNA Polymerase II Subunit A (Polr2a) as a highly expressed gene (middle column) and bacterial 4-hydroxy-tetrahydrodipicolinate reductase (DapB) as a negative control gene (right column). Histologic images show red puncta counterstained with hematoxylin. Each red puncta represents a single mRNA transcript [5]. Images were taken at 40x magni cation on an Aperio ImageScope CS2. Representative images of RNAscope assays using Mier3, Polr2a, and DapB probes on tissue sections of fresh FFPE rat mammary gland and carcinoma tissue. Tissues were formalin xed and processed immediately following resection (manufacturer recommended). Samples are from non-diseased mammary glands or DMBA-induced mammary carcinomas. RNAscope assay probes for Mier3 family member 3 (Mier3), as a moderately expressed gene (left column), RNA Polymerase II Subunit A (Polr2a), as a highly expressed gene (middle column), and bacterial 4-hydroxy-tetrahydrodipicolinate reductase (DapB), as a negative control gene (right column) were used. Histologic images show red puncta counterstained with hematoxylin. Each red puncta represents a single mRNA transcript [5]. Images were taken at 40x magni cation on an Aperio ImageScope CS Figure 5