Confirmation of canine acanthomatous ameloblastoma using RAS Q61R immunohistochemical staining of formalin-fixed paraffin-embedded tissues

Differentiating canine acanthomatous ameloblastoma (CAA) from oral squamous cell carcinoma (OSCC) based on routine histopathology can be challenging. We have previously shown that more than 95% of CAAs harbor an HRAS p.Q61R somatic mutation, while OSCCs carry either wild-type alleles or other MAPK pathway activating mutations (e.g., HRAS p.Q61L, BRAF p.V595E). Given that HRAS p.Q61R mutations are highly prevalent in CAA, we hypothesized that a RAS Q61R-specific rabbit monoclonal antibody may be a useful tool for confirmation of CAA by immunohistochemical (IHC) staining. In the present study, we assessed IHC staining of archived formalin-fixed and paraffin-embedded biopsy samples with a diagnosis of CAA (n = 23), using a RAS Q61R-specific rabbit monoclonal antibody (SP174) and an automated IHC stainer. Negative control samples consisted of HRAS p.Q61R mutation-negative OSCC tumors with either a known HRAS p.Q61L mutation (n = 1), BRAF p.V595E mutation (n = 4), or wild-type corresponding alleles (n = 3). We found that all 23 CAAs showed diffuse and strong membranous RAS Q61R immunoreactivity (100% sensitivity), while none of the 8 OSCCs showed immunoreactivity (100% specificity). The data supports the use of RAS Q61R-specific rabbit monoclonal antibody for diagnostic IHC confirmation of CAA and ruling out OSCC in dogs.


Introduction
Canine acanthomatous ameloblastoma (CAA) is a common neoplasm that originates from odontogenic epithelium (1)(2)(3)(4).It has been reported in multiple breeds of dogs of a wide age range (mean ~ 9 years) without an apparent sex predilection (5,6).The tumor is characterized by a discrete space-occupying mass arising from a dentated area, with a variable degree of invasion of underlying anatomical structures, including the jawbone (7,8).The current treatment of choice for CAA is wide-margin excision (i.e., mandibulectomy or maxillectomy), which typically offers long-term remission but is technically complex and frequently results in occlusal dysfunction (9)(10)(11)(12).Other treatment modalities have been described, including marginal excision, radiation therapy and intralesional bleomycin injections, but long-term remission rates are either low or variable, and side effects can be considerable (13)(14)(15).
Definitive diagnosis of CAA is based on routine histopathological examination (3,16).However, due to overlapping clinical, radiological, and histological features between CAA and oral squamous cell carcinoma (OSCC), up to ~30% of cases may be misdiagnosed (6,17).This has important clinical implications given fundamental differences in biological behavior including significantly more rapid growth as well as higher metastatic potential of OSCC relative to CAA (8,18).Although some of the molecular underpinnings driving OSCC and CAA tumorigenesis appear to be similar (e.g., aberrant MAPK pathway signaling) (19), we have shown in previous studies that their mutational landscape is distinct (20).Importantly, we have found that more than 95% of CAAs harbor an HRAS p.Q61R somatic mutation, while OSCCs typically harbor corresponding wild-type (WT) alleles or other MAPK pathway activating mutations (e.g., BRAF p.V595E, HRAS p.Q61L) (20,21).Arguably, identifying the mutational profile of tumor tissues is critical to distinguishing between these different epithelial cell tumors.However, genotyping experiments usually require molecular biology and/or sequencing capacity that are not readily available in standard diagnostic settings.Moreover, accurate diagnosis of CAA and differentiation from OSCC is critical to making informed clinical decisions and adequate prognostication early in the course of the disease.
Immunohistochemistry (IHC) is an alternative technique that does not require a highly specialized infrastructure and can be used to produce a signal that corresponds to a mutated protein, providing a means to infer point mutations (22,23).For this, monoclonal antibodies offering high sensitivity and specificity are required.One commonly used antibody, which we hypothesized might suit this purpose, is designated SP174.This recombinant rabbit monoclonal antibody was originally designed to detect human RAS Q61R, and is commercially-available (24).Of note, SP174 cross-reacts with different isoforms of RAS, making it useful for screening for p.Q61R mutations in NRAS, KRAS, and HRAS (22-26).Given that canine and human RAS isoforms share 98-100% sequence homology and based on the high frequency of HRAS p.Q61R somatic mutations in CAA (20,21,27), we hypothesized that SP174 could detect RAS Q61R in CAA and rule out OSCC.Therefore, the aim of this study was to determine the sensitivity and specificity of IHC using SP174 to detect RAS Q61R in CAA tumor tissue and a lack of immunoreactivity with OSCC, including tumors that harbor a closely related HRAS p.Q61L mutation.

Methods
Representative formalin-fixed and paraffin-embedded (FFPE) tissue sections from archived CAA and OSCC cases were assessed by IHC staining for the presence of RAS Q61R immunoreactivity using standard protocols and an automated IHC stainer (Bond-Max automated IHC staining system; Leica).All samples were submitted as biopsy specimens, obtained between 2013 and 2021, that did not undergo decalcification to preserve critical antigenic epitopes.The diagnosis of CAA was made based on initial hematoxylin and eosin assessment by a board-certified veterinary pathologist.The presence of an HRAS p.Q61R mutation was confirmed in the majority of CAAs using molecular assays, as previously described (8,20,21,28).Briefly, 4-μm tissue sections mounted on charged slides were deparaffinized (AR9222, Bond dewax solution; Leica), and after heat epitope retrieval (AR9640, Bond epitope retrieval solution 2; Leica) for 30-40 min, the slides were incubated with RAS Q61R-specific rabbit monoclonal antibody (SP174; ab227658; Abcam, Cambridge, MA, United States) raised against Ras mutated Q61R synthetic peptide (25) diluted 1:80 or 1:100 for 60 min followed by polymeric horseradish peroxidase (DS9800; Bond polymer refine detection; Leica) linker antibody conjugate detection system for 30 min, and hematoxylin counterstain (DS9390; Leica) for 5 min.Negative controls consisted of confirmed HRAS p.Q61R-negative OSCC FFPE tissues (20,21).Positive immunoreactivity was identified as brown staining with diaminobenzidine (DAB).The IHC evaluation was performed independently by two pathologists (M.M.M. and G.E.D), while blinded to the results of previous molecular analyses and medical records.
The use of archived diagnostic material, and/or review and data collection from medical records of client-owned animals for the purposes of this study was approved by Cornell University's Veterinary Clinical Studies Committee and was considered exempt from review by Cornell University's Institutional Animal Care and Use Committee.

Results
A total of 31 cases were enrolled representing 23 CAA and 8 OSCC tumors (Table 1; Figure 1).Of the 8 OSCC tumors, 5 corresponded to the papillary subtype (29).The average age was 8.85 ± 3.26 and 8.25 ± 3.11 for CAA and OSCC, respectively (p = 0.74, Mann-Whitney test).Sex distribution was 52% males and 48% with no significant differences based on tumor type (p = 0.92, Pearson's chi-squared test).Seventeen dog breeds were represented with Labrador retriever and mixed breed dogs being the two most common, accounting for 22.6 and 12.9%, respectively.Of the 31 cases, 17 CAA and all 8 OSCC had been previously genotyped using RNA-seq, PCR and Sanger sequencing and/or restriction fragment length polymorphism experiments (20,21).An HRAS p.Q61R somatic mutation was confirmed in all 17 genotyped CAA tumors.Only wild-type HRAS p.Q61 alleles were detected in 7 of the 8 OSCC tumors, and an HRAS p.Q61L mutation was detected in the remaining OSCC tumor.Four of the 8 OSCC tumors harbored a BRAF p.V595E mutation.

Discussion
In this study, we assessed the utility of RAS Q61R-specific SP174 antibody for confirmation of CAA and distinction from OSCC.Given fundamental biological differences (6,8,18), accurate distinction between CAA and OSCC is essential for proper clinical decision making and prognostication.Results showed that SP174 cross-reacts with RAS Q61R in CAA FFPE tissues with 100% sensitivity and 100% specificity, allowing accurate distinction of these relatively common tumor types.
Other IHC strategies have been previously proposed to differentiate between CAA and OSCC (31, 32).For example, distinction based on cytokeratin and calretinin protein expression patterns has been suggested (32), but assessment relies purely on semiquantitative scores and the diagnostic accuracy of these approaches has yet to be independently validated.Moreover, unbiased genomic profiling of CAA and OSCC shows variable expression patterns of the genes encoding these proteins without an obvious association with tumor type (19), suggesting that the patterns of cytokeratin and calretinin immunoreactivity in tumor tissues may not allow reliable tumor type distinction.
Another IHC-based strategy consists of determining the Ki67 labeling index of neoplastic epithelial cells (8,33,34).Ki67 is a nuclear protein expressed by cells that are actively engaged in the cell cycle (33).We have previously shown that the Ki67 labelling index of neoplastic cells is significantly lower in CAA compared to OSCC (8), which is a useful diagnostic complement to routine histopathological assessment.However, there is no established cutoff value that can reliably distinguish tumors with intermediate Ki67 labelling indices.In contrast, the strong and diffuse positive SP174 immunoreactivity of CAAs reported here offers a simple, easily interpreted binary distinction in which reactivity is either positive or negative, thus minimizing subjectivity, which is an inherent limitation with semi-quantitative scoring schemes commonly used for other IHC applications.Additionally, as shown in some of the cases in this report, CAA sections that contain non-neoplastic, normal surface gingival epithelium serve as an internal negative control minimizing interpretative bias.
Apart from CAA confirmation, the RAS Q61R-specific SP174 antibody may be used in the context of precision medicine, as proposed for human medical oncology (23)(24)(25)(26).Arguably, rapid detection of mutated genes using IHC allows attending clinicians to rationally select targeted therapies for neoplasms known to harbor HRAS, NRAS or KRAS p.Q61R mutations including ameloblastoma, melanoma, colorectal, urothelial and thyroid cancers, among others (35).A highly specific and comparable diagnostic approach is available using the monoclonal antibody 'VE1' for detection of BRAF p.V600E in human cancer (23,36,37).Given that homologous encoding somatic mutations (i.e., BRAF p.V595E) are highly recurrent in some tumors of dogs, including papillary OSCC and urothelial carcinoma (20,27,38), VE1 IHC might be a useful tool that would complement RAS Q61R-specific SP174 antibody IHC.However, the use of VE1 for detection of mutated BRAF protein in dogs has yet to be validated.Indeed, our attempts to optimize VE1 IHC protocols using OSCC tumors confirmed to harbor a BRAF p.V595E mutation have been unsuccessful despite a high degree of sequence homology (data not shown).Although a cause for the lack of VE1 antibody reactivity with the corresponding mutated protein in canine oral tumors is uncertain, we speculate that level expression of the mutated BRAF allele may be below the detection limit of IHC, which could account for the lack of reactivity.
As reported herein, reliable detection of RAS Q61R with SP174 antibody in tumor tissues is possible using FFPE tissues that have not  In conclusion, this study demonstrates that IHC with RAS Q61Rspecific SP174 antibody reagent is a highly sensitive, relatively simple, and specific method for confirmation of CAAs harboring an HRAS p.Q61R mutation.

FIGURE 1
FIGURE 1 Representative photomicrographs of canine acanthomatous ameloblastoma (CAA) case 9 and oral squamous cell carcinoma, papillary subtype (pOSCC), case 27 (Panels A,C, ×4; B,D, ×40 original magnifications).Low magnification of CAA (Panel A) showing characteristic anastomosing cords and trabeculae of odontogenic epithelium (arrows) within a background of stellate mesenchyme (asterisks).Higher magnification of a CAA (Panel C) showing neoplastic islands with characteristic outer layer of epithelial cells oriented with their long axes perpendicular to the basement membrane (palisading) and anti-basilar nuclear polarization (arrows).Low magnification of a pOSCC (Panel B) showing characteristic exophytic papillary pattern consisting of broad trabeculae of neoplastic stratified squamous epithelium (arrows) supported by fibrovascular connective tissue stroma (asterisks).Higher magnification of a pOSCC (Panel D) showing irregularly arranged epithelial cells that lack anti-basilar nuclear polarization and indistinct epithelial to mesenchymal junction at the periphery of individual neoplastic epithelial lobules (arrows).

FIGURE 2
FIGURE 2 Immunolabeling of canine acanthomatous ameloblastoma (CAA) with RAS Q61R-specific rabbit monoclonal antibody (SP174).Photomicrographs of representative CAA case 9 with a confirmed HRAS p.Q61R mutation.Note strong membranous RAS immunoreactivity throughout CAA neoplastic trabeculae and islands (Panels A, ×4 and B, ×40 original magnifications).The staining highlights anti-basilar nuclear polarization in CAA (Panel B).By contrast, a representative oral squamous cell carcinoma (OSCC), papillary subtype, case 27 with a confirmed HRAS p.Q61L mutation, shows a lack of RAS immunoreactivity throughout (Panels C, ×4 and D, ×40 original magnifications).In each panel, asterisks represent stroma and arrows indicate neoplastic epithelial cells arranged in anastomosing trabeculae and lobules.

TABLE 1
Summary case information.