RNA in situ hybridisation as a molecular diagnostic technique targeting IBA‐1 and CD204 in canine histiocytic sarcoma

Abstract Background Canine histiocytic sarcoma (HS) is an aggressive cancer with morphologically variable features; therefore, obtaining a definitive diagnosis can be challenging. Two proteins, IBA‐1, ionised calcium‐binding adapter molecule 1, and CD204, a macrophage scavenger receptor, have been shown to be specific immunohistochemical markers helpful in distinguishing HS from other tumour types with similar morphological features. Objectives This study was performed to demonstrate the use of RNA in situ hybridisation (ISH) technology allowing single‐molecule RNA visualisation in formalin‐fixed paraffin‐embedded (FFPE) tissues as a molecular tool for the diagnosis of canine HS. Methods Reverse transcription polymerase chain reaction (RT‐PCR) and western blot analysis for IBA‐1 and CD204 were performed to correlate gene expression and protein expression of these two markers in the histiocytic sarcoma DH82 cell line. RNA‐ISH for IBA‐1 and CD204 was performed on the DH82 cell line to validate the RNA‐ISH probes. RNA‐ISH and immunohistochemistry (IHC) were performed in clinical HS FFPE samples to demonstrate mRNA and protein expression of IBA‐1 and CD204. FFPE archived samples of canine round cell tumours, melanoma and anaplastic sarcoma were used as negative controls. Results RNA‐ISH and IHC showed moderate to strong expression for IBA‐1 and CD204 in the neoplastic cells in both the canine DH82 cell line and the archived canine HS samples. RNA‐ISH and IHC showed scattered positive staining in the control tumours samples, consistent with macrophagic infiltration. Conclusion RNA‐ISH for CD204 and IBA‐1 appeared to have a high specificity and sensitivity in our samples and may be an additional valuable diagnostic technique in identifying HS.


INTRODUCTION
Canine histiocytic sarcoma (HS) is a highly aggressive neoplasm. It can be characterised into three forms: localised, disseminated and haemophagocytic. Both localised and disseminated HS are derived from an interstitial dendritic cell origin and are thought to be separate stages of the same disease (Affolter & Moore, 2002;Fulmer & Mauldin, 2007;Moore et al., 2006). Haemophagocytic HS is of a macrophagic origin. All three types of HS are derived from the same haematopoietic stem cell precursors (Affolter & Moore, 2002).
HS is difficult to diagnose as its morphological features are highly variable (Affolter & Moore, 2002;Erich et al., 2018;Wakahashi et al., 2010). The morphologic features of HS can often mimic other tumours of mesenchymal or round cell origin such as sarcomas, plasma cell tumour, mast cell tumour, lymphoma, transmissible venereal tumour or amelanotic melanoma. In Erich et al,47.5% of cases diagnosed with histiocytic sarcoma or its differential diagnoses using histology were reclassified using morphological criteria paired with an immunohistochemistry (IHC) protocol (Erich et al., 2018).
In the past decade, CD204 has been described as a useful marker for identification of cells of macrophagic origin and as a more specific immunohistochemical marker for canine HS (Kato et al., 2014;Kato et al., 2013;Ide et al., 2011;Thongtharb et al., 2016). CD204, or class A macrophage scavenger receptor, is a receptor typically expressed in macrophages. In Kato et al. (2013), CD204 was found to be highly expressed in all HS samples, including those of both macrophagic and dendritic phenotypes. CD204 was not expressed by other round cell tumours or sarcoma samples in this study, except in one grade III mast cell tumour, apart from infiltrating macrophages. In Ide et al, most pleomorphic neoplastic cells in subdural HS expressed CD204 (Ide et al., 2011). Similarly, in Thongtharb et al. (2016), all tumour samples (primary intracranial HS) were found to have a moderate to strong expression of CD204.
Additionally, IBA-1, or ionised calcium-binding adapter molecule 1, a protein mediating calcium signals in monocytic and histiocytic cells, has been reported to be a specific diagnostic marker for canine HS (Ide et al., 2011). It is expressed in all cells of macrophage or monocyte origin, including all three forms of HS (Pierezan et al., 2014). In one study, selected cases of histiocytic disease, including canine histiocytomas, reactive histiocytosis and histiocytic sarcomas, were found to have a moderate to strong cytoplasmic immunoreactivity for IBA-1, whereas samples of melanomas, lymphomas, mast cell tumours and plasma cell tumours were negative, apart from scattered positive non-neoplastic stromal and peritumoural histiocytes as well as epidermal Langerhans cells (Pierezan et al., 2014). Similarly, all HS tumour samples assessed by Thongtharb et al. (2016) and by Ide et al. (2011) showed a moderate to strong immunoreactivity for IBA-1.
Therefore, both IBA-1 and CD204 are considered to be useful, specific markers in the diagnosis of HS by distinguishing it from similar conditions (Kato et al., 2013;Kato et al., 2014;Pierezan et al., 2014).
RNAscope® in situ Hybridisation (ISH) is a technique to evaluate mRNA expression. It allows for visualisation of a single molecule of RNA in both formalin fixed tissues and cytological samples (Bingham et al., 2017;Chan et al., 2018). This technology allows for analysis of mRNA expression without destroying tissue architecture.
The primary goal of this study was to evaluate the expression of IBA- antibiotic-antimycotic (Gibico) at 37 • C in 5% CO 2 until 90-100% confluency.

Cell pellets
Cell pellets, derived from the DH82 cell line, were utilised for both immunohistochemistry (IHC) and RNA-ISH staining. Once DH82 reached 90-100% confluence, cells were trypsinised, transferred into 50 ml conical tubes and then centrifuged at 400 g for 10 min. Cell pellets were fixed in 10% neutral buffered formalin, and paraffin embedded pellets were prepared using histogel-based cell block preparation (HISTOGEL, ThermoFisher Scientific, Waltham, MA) following manufacturer's instructions.

Formalin-fixed paraffin-embedded (FFPE) tissue
Archived paraffin-embedded blocks of 8 canine HS and 10 canine tumour types commonly included in the differential diagnosis (DDx) of HS (oral malignant melanoma n = 2, mast cell tumour n = 2, lymphoma n = 2, plasma cell tumour n = 2 and anaplastic sarcoma n = 2) were obtained. Histologic sections of tumour were reviewed, and a definitive diagnosis was made by board-certified veterinary pathologists at one public academic institution. A second review was performed by one single board-certified veterinary pathologist (M,D) to confirm the diagnosis. The tissues had been fixed in 10% buffered neutral TA B L E 1 Primer sequence and predicted product size
The sequences of primers and predicted size applications are listed in Table 1. RT-PCR was performed by combining 19 µl of RNAse free water, 25 µl of master mix (HotStar Taq, Qiagen, Germantown, MD), 2 µl of cDNA generated previously, 2 µl of the forward primer and 2 µl of the reverse primer. The samples were placed in the thermocycler and incubated at 94 • C for 5 min, followed by 35 cycles consisting of 94 • C for 45 s, 57 • C for 45 s, 72 • C for 45 s, followed by 5 min of 72 • C in order to denature, anneal and extend the sample, respectively. The PCR products were run on 1% agarose gel electrophoresis with 0.01% ethidium bromide. The gel was analysed using a gel imager (FluorChemE, Cell Biosciences, San Jose, CA). The bands of the PCR products were cut out under UV light exposure, and the cDNA products were isolated using a commercially available kit (PureLink™ Quick Gel Extraction Kit, ThermoFisher Scientific) following manufacturer instruction. The purified cDNAs were submitted for Sanger sequence analysis (GENEWIZ; South Plainfield, NJ, USA) to compare canine genetic sequences at the GenBank.

Western blot
Western blot was performed on the DH82 cell line to validate the pro-

Immunohistochemistry
IHC was performed on the DH82 cell pellets and on the FFPE clinical tissues using the automated Leica BOND RX platform. Briefly, the slides were heated at 60 • C for 10 min, then deparaffinised on the BOND RX using Dewax (Leica Biosystems, Buffalo Grove, IL, USA). The solution is heated to 72 • C for 30 s. The slides are then rinsed with three changes of 100% reagent alcohol. They were rinsed in Bond Wash for 3 min, and then rinsed with DI water for 4 min. The slides were then treated with ER1 (antigen retrieval buffer) at 100 • C for 20 min.
The samples were treated with hydrogen peroxide and incubated for 10 min, and then rinsed with Bond Wash three times, for 2 min each.
The slides were treated with identical mouse monoclonal primary antibody for either IBA-1 (1:300, MiliporeSigma) or CD204 (1:200, Transgenic Inc), which was used in western blot above, and incubated for 15 min (Kato et al., 2013;Pierezan et al., 2014).  Five randomly selected areas were counted for assessment with 40× magnification, following manufacturer's recommendation. High cellularity areas were considered as proper evaluation sites, and if significant cellular effects such as necrosis or haemorrhage were seen, they were avoided for the evaluation. In this study, we defined all HS samples as positive in our methods if they were scored as a 1, 2 or 3.

STATISTICS
In order to evaluate the mRNA expression of IBA-1 and CD204 on 8

mRNA expression of IBA-1 and CD204 in DH82 cell line
RT-PCR was performed to confirm the mRNA expression of IBA-1 and CD204 in the DH82 cell line. Bands were detected for both target genes ( Figure 1). GAPDH, the reference gene, was also detected in the DH82 cell line. The PCR products were sequenced, and canine specific IBA-1 and CD204 expression was confirmed. IHC was performed on archived 8 HS and 10 other selected tumour types, as described above, to verify and visualise the expression of IBA-1 and CD204. The targets were visualised using a brown chromogen.
The majority of the HS samples showed both positive cytoplasmic and membrane staining pattern and high intensity staining with IBA-1 and CD204 (Figure 3a and b). IBA-1 showed a score 3 on all 8 HS samples.
CD204 showed a score 3 on 6 HS samples and a score 2 on 1 HS sample, but 1 sample did not show positivity on CD204 (Table 3). In the samples of lymphomas, mast cell tumours, melanomas, plasma cell tumours and anaplastic sarcomas, immunoreactivity for IBA-1 or CD204 was absent in the neoplastic cells. Scattered positive non-neoplastic macrophages were noted infiltrating the tumour stroma (Figure 3c and d).

4.4
In situ expression of mRNA for IBA-1 and CD204 genes IBA-1 and CD204 mRNA expression was assessed in the DH82 cell line pellets via the RNA-ISH method to validate both RNAscope® probes (Supplemental Data S3A and B).
RNA-ISH was also performed on 8 HS and 10 other selected tumour types, as described above. IBA-1 and CD204 mRNA dots were detected in all 8 HS samples (Figure 4a and b). They were clearly visualised as punctate red dots, suggesting that IBA-1 and CD204 are strong markers for canine HS. On the other hand, little and scattered staining was seen in all other tumour types, suspected to be macrophage infiltration (Figure 4c and d). The IBA-1 and CD204 mRNA expression was evaluated using the ACD RNAscope® ISH scoring system as a semiquantitative gene expression analysis. Relative target gene expression was evaluated by manually counting the number of clear dots per cell.
Most FFPE HS clinical samples showed a high level of IBA-1 mRNA expression (score 4 n = 3, score 3 n = 4 and score 2 n = 1). Similarly, CD204 mRNA was found to have a high level of expression in the HS samples (score 4 n = 4, score 3 n = 3 and score 2 n = 1) (Figure 5, F I G U R E 3 Immunohistochemistry on HS and mast cell tumour (common differential diagnosis: DDx). Cytoplasmic and membrane staining were seen using with IBA-1 (a and c) and CD204 (b and d) antibodies. High intensity and more than 70% staining area were seen on this HS sample (a and b, score 3). On the other hand, scattered staining was seen on this mast cell tumour, which suggested macrophage infiltration (c and d). Scale bar indicates 50 µm Table 3). No significant scoring difference was seen between IBA-1 and CD204 expression in HS.

DISCUSSION
The present study reports the validation of an in situ analysis of IBA-1 and CD204 mRNA expression in retrospective FFPE samples of canine histiocytic sarcoma. In this project, we elected to apply a semiquantitative RNA-ISH method to assess expression of these markers in the neoplastic cells. We found that expression of the mRNAs for IBA-1 and CD204 was high in most HS samples. In addition, a strong correlation was observed between the IHC score and the RNA-ISH results. None of the neoplastic cells in our control tumours showed immunoreactivity for CD204 or IBA-1, confirming that IBA-1 and CD204 are additional useful markers in the diagnosis of canine HS.
Scattered non-neoplastic cells were observed to show IBA-1 or CD204 protein or mRNA expression in the non-HS tumour samples, likely consistent with a macrophagic infiltration. These markers' specificity for cells of the monocyte/macrophage lineage is in correlation with previous published reports (Ide et al., 2011;Pierezan et al., 2014;Thongtharb et al., 2016). One histiocytic FFPE HS sample did not show positive immunoreactivity for CD204 during our IHC analysis but was positive on IHC for IBA-1. Positive and negative IHC controls showed adequate results. The same FFPE HS sample did show positive labelling for CD204 mRNA upon RNA-ISH. We suspect that the negative IHC labelling observed could be related to the IHC protocol and processing (Ramos-Vara, 2005). It is also possible that, in this HS tumour sample, the CD204 protein may not be detectable with the IHC antibodies used due to post translational modifications or aberrant protein processing (Liu et al., 2016;Qin et al., 2020;Van Drie, 2011).
RNAscope® is compatible with routine formalin-fixed, paraffinembedded tissue specimens and can allow for multiplex analysis. In addition, RNAscope® analysis for IBA-1 or CD204 can be applied in cytological samples, which may be useful in the clinical setting as it may provide a definitive diagnosis of HS using a small sample such as a fine needle aspirate (Shiomitsu et al., 2020).

F I G U R E 4
The mRNA expression of IBA-1 and CD204 in archived HS (a and b) and plasma cell tumour (c and d). The samples probed with IBA-1 (a and c) and CD204 (b and d) displayed high mRNA expression on HS. RNA-ISH of IBA-1 and CD204 showed 10-15 dots/cell and/or <10% dots are in clusters (ACD score: 3). On the other hand, there is minimum scattered staining on plasma cell tumour (common differential diagnosis: DDx). Scale bar indicates 50 µm The RNA ISH method brings the benefits of in situ analysis and quantification of RNA biomarkers and can be supplemental to existing IHC methods if no antibodies are available (Wang et al., 2012). In our study, we elected to assess mRNA expression applying a manual count of the dots visualised through RNA ISH. We chose to use the suggested semi-quantitative scoring system in which the number of dots per cells is recorded and associated with a specific score, as long as the positive control is expressed in adequate level (A Guide for RNAscope Data Analysis: Advanced Cell Diagnosis Manual, 2021). Our HS samples showed high scores of 2 and 3 using this method. Quantification of IBA-1 and or CD204 mRNA expression via RNA-ISH might be useful for the prediction of treatment response or prognosis, and we intend to assess expression scores with outcomes as we expand our selection of FFPE HS samples. A relative quantification could be applied utilising a dual detection RNA-ISH method to compare the target gene expression with reference genes (Wolff et al., 2018). Manual enumeration of dots can be cumbersome; in the future, we intend to employ an automated measurement protocol using software analysis such as HALO software (Indica Labs) in order to improve the analytic performance and reduce human subjectivity in recognising the dots.
Due to the nature of our study (validation study), only a small number of FFPE tumour types were selected to act as possible negative samples. We selected cases of canine lymphomas, mast cell tumours, melanoma, plasma cell tumours and anaplastic sarcomas as these neoplastic processes can sometimes be misdiagnosed as HS. Recruiting additional tumour samples would help to confirm the specificity of the probes used. Another limitation would be unknown inter-observer variability because of a single pathologist evaluation. It would be beneficial to involve more than 1 pathologist to minimise a potential inconsistency of the histopathological and immunohistological assessment.

CONCLUSION
While IHC remains a gold standard and widespread technique used in the diagnosis of canine HS, RNA-ISH technology could be an additional F I G U R E 5 IHC and RNA-ISH Scoring outcomes on 8 HS and 10 other types of tumours including common differential diagnosis (DDx) of HS archived samples. No significance staining score deference was seen IBA-1 and CD204 mRNA expression with RNA-ISH on 8 HS (IBA-1: median was 3, CD204: median was 3.5). High IHC score and high staining RNA-ISH score were seen on 8 HS with both IBA-1 and CD204. Low IHC score and low staining RNA-ISH score were seen on 10 other types of tumours including common DDx with both IBA-1 and CD204 useful diagnostic tool for identifying interested genes. This is the first study utilising RNA in situ hybridisation to identify distinguishing targets, IBA-1 and CD204, to allow for the diagnosis of canine HS. We were able to validate the use of RNA-ISH targeting IBA-1 and CD204 in FFPE HS samples selected retrospectively and in our HS DH82 cell pellets. This molecular diagnostic technique can be paired with IHC to allow analysis of mRNA and protein, respectively, while also preserving the cellular relationship and tissue architecture.