A PIK3CA transgenic mouse model with chemical carcinogen exposure mimics human oral tongue tumorigenesis

Summary

Oral cancer causes significant global mortality and has a five‐year survival rate of around 64%. Poor prognosis results from late‐stage diagnosis, highlighting an important need to develop better approaches to detect oral premalignant lesions (OPLs) and identify which OPLs are at highest risk of progression to oral squamous cell carcinoma (OSCC). An appropriate animal model that reflects the genetic, histologic, immunologic, molecular and gross visual features of human OSCC would aid in the development and evaluation of early detection and risk assessment strategies. Here, we present an experimental PIK3CA + 4NQO transgenic mouse model of oral carcinogenesis that combines the PIK3CA oncogene mutation with oral exposure to the chemical carcinogen 4NQO, an alternate experimental transgenic mouse model with PIK3CA as well as E6 and E7 mutations, and an existing wild‐type mouse model based on oral exposure to 4NQO alone. We compare changes in dorsal and ventral tongue gross visual appearance, histologic features and molecular biomarker expression over a time course of carcinogenesis. Both transgenic models exhibit cytological and architectural features of dysplasia that mimic human disease and exhibit slightly increased staining for Ki‐67, a cell proliferation marker. The PIK3CA + 4NQO model additionally exhibits consistent lymphocytic infiltration, presents with prominent dorsal and ventral tongue tumours, and develops cancer quickly relative to the other models. Thus, the PIK3CA + 4NQO model recapitulates the multistep genetic model of human oral carcinogenesis and host immune response in carcinogen‐induced tongue cancer, making it a useful resource for future OSCC studies.

1. INTRODUCTION

Oral squamous cell carcinoma (OSCC) has a poor five‐year survival rate of around 64%, primarily due to diagnosis at a late stage. ¹ , ² Risk factors for OSCC include exposure to carcinogens from tobacco and alcohol use, environmental factors and inherited genetic traits. ³ Most oral cancers are preceded by oral premalignant lesions (OPLs), a group of clinically suspicious conditions including leukoplakia (white plaques), erythroplakia (red plaques) and submucous fibrosis that can eventually transform to OSCC if left untreated. ⁴ OPLs exhibit variations in colour, surface texture and thickness that can be observed during a conventional oral examination. ^{5 , 6} However, because it is difficult to distinguish between OPLs with low risk and high risk for cancer progression, OPLs remain suboptimally managed under the standard of care. ⁷ The literature suggests that certain OPL biomarkers, including changes in gross visual appearance, histologic evidence of dysplasia, immune response, molecular biomarker expression and genetic alterations, could all be informative in assessing the risk of progression to OSCC. ⁵ , ⁶ , ⁸

Dysplasia is often observed in OPLs with mild, moderate and severe dysplasia grades defined according to the depth and severity of the changes. ⁹ , ¹⁰ OPLs with dysplasia have a higher likelihood of malignant transformation to OSCC, with increasing risk accompanying more severe grades of dysplasia. ⁹ , ¹⁰ Dysplasia in OPLs is sometimes accompanied by an immune cell infiltrate composed of mostly lymphocytes within the underlying lamina propria. Lymphocytic infiltration is involved in cancer immunosurveillance, where lymphocytes recognize and eliminate precancerous and cancerous cells with abnormal antigen expression. ¹¹ Studies show that the presence of lymphocytic infiltration in dysplastic OPLs has been associated with limited progression to OSCC and favourable treatment outcomes. ¹² , ¹³

Multiple molecular biomarkers have been identified as potentially indicative of OPL risk, though most have not been validated in large prospective studies. ¹⁴ Elevated expression of the cell proliferation marker Ki‐67, which has been identified in prospective cohort analysis, has been linked to the malignant transformation risk of leukoplakias and locoregional recurrence following OSCC treatment. ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ Loss of heterozygosity is another biomarker identified in prospective cohorts as predictive of OPL progression risk. ¹⁹ , ²⁰ Additionally, some studies suggest that inflammation is associated with cancer progression and could have a role in promoting malignant transformation. ²¹

Despite the potential for these biomarkers to be incorporated into the screening or diagnostic process for oral cancer, a conclusive set of indicators with strong prognostic significance has yet to be identified. ¹⁸ The availability of animal models with biomarkers mimicking those found in human oral carcinogenesis could greatly facilitate studies to understand oral cancer progression and evaluate better methods for OPL risk assessment.

There are established animal models of OSCC where carcinogenesis is induced by exposure to a chemical carcinogen. ²² One model is the 7,12‐dimethylbenz[a]anthracene (DMBA) hamster cheek model; however, this model is not reflective of human OSCC in that the hamster cheek lacks lymphatic drainage and the immune response differs from the T‐cell response expected in human OSCC. ²³ , ²⁴ , ²⁵ Another established model is the 4‐nitroquinoline 1‐oxide (4NQO) mouse tongue model, in which wild‐type mice are exposed to the chemical carcinogen 4NQO, a DNA adduct‐forming agent that produces genetic and epigenetic alterations similar to that of tobacco‐derived carcinogens. ²⁶ This model can take almost one year to develop, as it requires about 16 weeks of 4NQO exposure and then an additional 12 to 33 weeks before OSCC is observed. ²² , ²⁴

More recently, researchers have developed transgenic mouse models of OSCC that reflect genetic changes found in human OSCC. In these transgenic models, oncogene activation and/or tumour suppressor inactivation in transgenic models initiates carcinogenesis without requiring exposure to chemical carcinogens. Specific approaches employed in transgenic models of OSCC include overexpression of the K‐ras oncogene and the loss of phosphatase and tensin homolog tumour suppressor gene. ²² Some transgenic models also incorporate exposure to chemical carcinogens, typically 4NQO. The combination of chemical carcinogen and genetic influences mirrors a human with genetic or epigenetic OSCC predispositions being exposed to tobacco, alcohol or other carcinogens. ²²

The oncogene PIK3CA is involved in the activation of the PI3K/AKT/mTOR pathway that is implicated in multiple types of human cancers, and is one of the more commonly mutated genes in oral cancers. ²⁷ , ²⁸ , ²⁹ , ³⁰ We present our experimental genetically engineered PIK3CA + 4NQO and PIK3CA‐E6/E7 mouse models of oral tongue carcinogenesis and compare temporal changes in histologic, immunologic, molecular and gross visual biomarkers of these experimental models to an existing model in which wild‐type CBA mice are exposed to 4NQO (referred to as the CBA/Ca + 4NQO model).

2. METHODS

2.1. Ethical approval statement

Animal care and experimental procedures were performed under a protocol reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Texas School of Dentistry at Houston.

2.2. Animals

2.2.1. Model 1. CBA/Ca + 4NQO

The established CBA/Ca + 4NQO model involves chemical carcinogen‐induced oral carcinogenesis in an inbred mouse strain and was used as a reference for comparison against the transgenic models. To generate this OSCC model, CBA/Ca mice (Jackson Laboratory, Bar Harbor, ME) were exposed to 4NQO (Sigma‐Aldrich, St. Louis, MO) in their drinking water at a concentration of 100 μg/mL for a maximum duration of 16 weeks Table 1.

TABLE 1.

An existing CBA/Ca + 4NQO mouse model of chemical carcinogen‐induced oral carcinogenesis, an experimental transgenic PIK3CA + 4NQO model with both chemical carcinogen and genetic influences of carcinogenesis, and an experimental transgenic PIK3CA‐E6/E7 model with only genetic influences of carcinogenesis were compared. Table indicates the duration of exposure to the chemical carcinogen and/or oncogene activators

		Chemical carcinogen	Oncogene activators
		4NQO	Tamoxifen	Doxycycline
Existing model	(1) CBA/Ca + 4NQO	16 wk	—	—
Experimental transgenic models	(2) PIK3CA + 4NQO	30 d	5 d	—
	(3) PIK3CA‐E6/E7	—	5 d	60 d

2.2.2. Transgenic models

Both experimental genetically engineered models originated from K14‐CreERTM/LSL‐rtTA/Tet‐E6/E7 (heterozygous male) X K14‐CreERTM/LSL‐PIK3CAH1047R (heterozygous female) mice (provided by Dr Silvio Gutkind, University of California San Diego, CA). Breeding generated offspring with the genotypes K14‐CreERTM/LSL‐rtTA/LSL‐PIK3CAH1047R (genotype referred to as PIK3CA) and K14‐CreERTM/LSL‐rtTA/LSL‐PIK3CAH1047R/tet‐E67 (genotype referred to as PIK3CA‐E6/E7). The genotype of each transgenic offspring was assessed using tissue collected from a 2‐mm tail biopsy 18 days after birth according to standard protocols.

2.2.3. Model 2. PIK3CA + 4NQO

The PIK3CA + 4NQO transgenic mouse model was designed to explore the combined effects of genetic and chemical carcinogen influences of oral carcinogenesis. It contains the PIK3CA c.3140A > G (H1047R) hotspot mutation in the heterozygous state which, following exposure to tamoxifen, overexpresses Cre recombinase‐mediated expression mutant PIK3CA +/‐ in the oral epithelium. The PIK3CA mutation promotes oral carcinogenesis by constitutively activating the PI3K pathway resulting in sustained proliferative signalling, evasion of growth suppression and genomic instability. ⁸ Mice were exposed to tamoxifen (50 μL at 200 mg/mL dissolved in DMSO, Sigma‐Aldrich, St. Louis, MO) applied topically to their tongues for the first 5 days of the study to activate the mutant PIK3CA oncogene. DNA adduct‐forming 4NQO, a surrogate carcinogen for tobacco exposure, was delivered at a concentration of 100 μg/mL in their drinking water for the first 30 days of the study Table 1.

2.2.4. Model 3. PIK3CA‐E6/E7

The PIK3CA‐E6/E7 transgenic mouse model was designed for spontaneous oral carcinogenesis as a result of oncogene activation alone. In addition to the mutant PIK3CA oncogene, high‐risk human papillomavirus (HPV)‐derived E6 and E7 oncogenes promote the development of HPV‐associated OSCC by inhibiting the p53 and retinoblastoma tumour suppressor genes, respectively. ⁸ This model uses ROSA26‐rtTA‐IRES‐EGFP, a lox‐stop‐lox system with a reverse tetracycline transactivator (rtTA), to concurrently express E6 and E7 oncogenes and PIK3CA in the oral epithelium. This model conditionally expresses E6 and E7 genes upon rtTA activation by doxycycline, a tetracycline derivative. Mice were exposed to tamoxifen (50 μL at 200 mg/mL dissolved in DMSO, Sigma‐Aldrich, St. Louis, MO) applied topically to their tongues for the first 5 days of the study for PIK3CA activation. Concurrently, mice were exposed to doxycycline, a tetracycline derivative, through grain‐based pellets (6 g/kg, Envigo, Madison, WI) in their diet for the first 60 days of the study to modulate E6 and E7 expression Table 1.

2.3. Experimental procedures

Mice were euthanized using carbon dioxide according to standard protocols. One mouse from each model was sacrificed at the beginning of the study to serve as a baseline. Subsequently, mice from each model were sacrificed within the time frames of 3‐8, 9‐12, 13‐16 and 17‐20 weeks after initiation of exposure to 4NQO, tamoxifen and/or doxycycline. The number of mice from each model is shown in Table 2 categorized by sacrifice time frame and pathology diagnosis. As a result of compliance with IACUC humane endpoint guidelines, there were no PIK3CA + 4NQO or PIK3CA‐E6/E7 mice remaining after 20 weeks. Two mice from the CBA/Ca + 4NQO model were sacrificed in the 21 + week time frame.

TABLE 2.

Number of mice from each model sacrificed over the course of the study categorized by sacrifice time frame and pathology diagnosis

	Weeks	No dysplasia	Mild dysplasia	Moderate dysplasia	Severe dysplasia	OSCC	Total
(1) CBA/Ca + 4NQO	3‐8	0	4	0	0	0	4
	9‐12	0	0	3	2	0	5
	13‐16	0	0	7	0	0	7
	17‐20	0	1	1	9	0	11
	21+	0	0	0	0	2	2
	n = 29
(2) PIK3CA + 4NQO	3‐8	1	2	2	1	0	6
	9‐12	0	0	0	1	1	2
	13‐16	0	0	0	0	1	1
	17‐20	0	0	1	0	1	2
	n = 11
(3) PIK3CA‐E6/E7	3‐8	2	0	3	0	0	5
	9‐12	0	0	1	2	0	3
	13‐16	0	0	1	1	0	2
	17‐20	0	0	1	0	0	1
	n = 11

2.4. Gross visual appearance

After sacrifice, each mouse tongue was carefully dissected from the mouth and rinsed in 1X PBS to remove excess blood. Dissected tongues underwent gross examination under magnifying glass by an oral and maxillofacial pathologist (NV) to identify grossly visible lesions. Images of the fresh tissue specimens were captured using a Nikon D700 camera with a Nikkor 105‐mm macro lens (Nikon, Tokyo, Japan).

2.5. Histologic analysis

The whole tongues were longitudinally bisected or serially sectioned to include grossly visible lesions for histopathologic examination. Sections were fixed in 10% neutral buffered formalin, embedded in paraffin, sliced to 4 µm and stained with haematoxylin and eosin (H&E). Six serial H&E‐stained tissue sections, three from each side of the bisected margin, were assessed from each specimen. Dorsal and ventral surfaces of the tongue were evaluated for dysplasia grade and in situ and invasive squamous cell carcinoma by the study pathologist (NV) using established criteria for grading oral epithelial dysplasia. Briefly, mild dysplasia was defined as hyperplasia of basal and parabasal cells with minimal cytological atypia and architectural changes confined within the lower third of the epithelium; moderate dysplasia as cytological atypia and architectural changes extending into the middle one‐third of the epithelium; and severe dysplasia as prominent cytological atypia and marked architectural changes extending into the upper third of the epithelium. ⁹ H&E‐stained sections were additionally evaluated for the presence of lymphocytic infiltrate within the lamina propria in close proximity to the epithelial cells. The presence or absence of lymphocytic infiltration was assessed at the site of most severe dysplasia on the dorsal tongue and the site of most severe dysplasia on the ventral tongue for each specimen.

2.6. Immunohistochemistry for molecular analysis

Ki‐67 was selected as a candidate molecular biomarker for immunohistochemical analysis based on literature reporting its association with oral carcinogenesis and OPL progression risk. ¹⁸ Immunohistochemical staining was performed on 4‐μm formalin‐fixed paraffin‐embedded sections. Staining for Ki‐67 (monoclonal, rabbit, 1:50, Cell Signaling Technology, Danvers, MA) was performed according to the manufacturer's protocol. One baseline specimen and one specimen from an intermediate time point (14‐18 weeks) were stained from each model. Intermediate time points were evaluated because they had dysplasia with potential for malignant transformation, unlike later time points where OSCC had already developed. The presence and extent of Ki‐67 immunohistochemical staining were qualitatively assessed by the study pathologist (NV).

3. RESULTS

3.1. Gross visual appearance

Distinctive tongue surface abnormalities, including tumours and raised white plaques resembling leukoplakias in human OSCC, were observed in the PIK3CA + 4NQO model throughout the study. Figure 1 shows representative gross photographs of ex vivo tongue specimens from an intermediate time point for each model, where paired dorsal and ventral tongue images are from different sides of the same specimen. Slight surface texture irregularities in the dorsal tongue were observed in the CBA/Ca + 4NQO model, and no abnormalities were observed in the ventral tongue. The PIK3CA + 4NQO model had raised white plaques and prominent tumours on the dorsal and ventral tongue, while the PIK3CA‐E6/E7 model did not exhibit either dorsal or ventral tongue surface abnormalities.

FIGURE 1.

Representative gross photographs of ex vivo tongue specimens. The PIK3CA + 4NQO model exhibited the most prominent white plaques and tumours on the dorsal and ventral tongue. For scale, mouse tongue length is reported in the literature as approximately 1 cm ³⁵

3.2. Histologic analysis

Figure 2 shows representative dorsal tongue histopathology sections of specimens with mild or moderate dysplasia, severe dysplasia and OSCC for each model, where applicable. OSCC was not observed in the PIK3CA‐E6/E7 model during the study. Each mouse model exhibited microscopic features that resembled human tongue carcinogenesis. Figure 3 shows representative ventral tongue histopathology sections from each model at the 9‐week time point. Mild to moderate dysplasia in the ventral tongue was observed in the transgenic PIK3CA + 4NQO and PIK3CA‐E6/E7 models but was not observed in the CBA/Ca + 4NQO model. The PIK3CA‐E6/E7 model was also discovered to have uniquely developed tumours in the lips and perioral region (Figure S1).

FIGURE 2.

Histopathology sections from dorsal tongue specimens over the time course show mild or moderate dysplasia, severe dysplasia and OSCC from each model, where applicable. Cytological and architectural features are similar to those observed in human tongue carcinogenesis. Scale bars represent 100 μm

FIGURE 3.

Histopathology sections of representative ventral tongue specimens at 9 wks. No dysplasia was observed in the CBA/Ca + 4NQO model, but mild to moderate dysplasia was present in the PIK3CA + 4NQO and PIK3CA‐E6/E7 models. Scale bars represent 100 μm

Overall, OSCC was observed earliest in the time course and most frequently in the PIK3CA + 4NQO model. Comparing the proportion of specimens that developed OSCC before the end of 20 weeks using the Freeman‐Halton extension of Fisher's exact test yielded P = .009, indicating that the three models are not equally likely to develop OSCC by this time point. Figure 4 shows the cumulative total of specimens with severe dysplasia or OSCC as a percentage of the cumulative total of mice sacrificed up to that point. Histologic analysis showed that the first observed instances of severe dysplasia in the CBA/Ca + 4NQO model occurred in the 9‐ to 12‐week time frame. In contrast, the first observed instances of severe dysplasia in both transgenic PIK3CA + 4NQO and PIK3CA‐E6/E7 models occurred earlier in the 3‐ to 8‐week time frame. OSCC was only observed in the CBA/Ca + 4NQO model in the 21 + week time frame. However, in the PIK3CA + 4NQO model, OSCC was first observed during the 9‐ to 12‐week time frame. Subsequent time frames for the PIK3CA + 4NQO model showed increases in OSCC incidence as a percentage of total sites assessed.

FIGURE 4.

Percentage of severe dysplasia (grey) and OSCC (black) specimens observed in each model throughout the study. The cumulative number of mice sacrificed up to and including each time frame is indicated by n. OSCC was observed in the transgenic PIK3CA + 4NQO model earlier than the other models

Evaluation of histopathology sections identified foci of lymphocytic infiltrate within the lamina propria, as shown in representative moderate dysplasia and severe dysplasia histopathology sections from the PIK3CA + 4NQO model in Figure 5A. While lymphocytic infiltrate was observed in some specimens from each model, it was most frequently and consistently observed in the PIK3CA + 4NQO model. The number of sites with lymphocytic infiltrate as a percentage of sites assessed is shown in Figure 5B, classified by dysplasia grade and model. Lymphocytic infiltrate was observed in a subset of sites with moderate and severe dysplasia in the CBA/Ca + 4NQO model. In the PIK3CA + 4NQO model, lymphocytic infiltrate consistently presented in at least 60% of sites across all dysplasia grades. However, the lymphocytic infiltrate was only observed in a few PIK3CA‐E6/E7 specimens with severe dysplasia. In these specimens, lymphocytic infiltration was not observed at either normal or OSCC sites.

FIGURE 5.

A, Representative histopathology sections showing lymphocytic infiltrate in the PIK3CA + 4NQO model, indicated by black arrows. Scale bars represent 100 μm. B, Percentage of tongue sites assessed that exhibited lymphocytic infiltrate. Number of sites assessed in each pathology category indicated by s. Lymphocytic infiltrate was most consistently observed in the PIK3CA + 4NQO model across all dysplasia grades

3.3. Immunohistochemistry for molecular analysis

Figure 6 shows representative dorsal and ventral tongue sites with Ki‐67 staining at baseline and at an intermediate time point for each model. Overall, Ki‐67 expression was relatively unchanged between baseline and the intermediate time point in the CBA/Ca + 4NQO model, while it slightly increased in the PIK3CA + 4NQO and PIK3CA‐E6/E7 models. All dorsal tongue sites at the intermediate time point (CBA/Ca + 4NQO 18 weeks; PIK3CA + 4NQO 17 weeks; PIK3CA‐E6/E7 14 weeks) had the same histopathology grade of severe dysplasia. The transgenic models showed a greater increase in Ki‐67 staining between baseline and the intermediate time point than the CBA/Ca + 4NQO model. The ventral tongue sites on the same specimen had different dysplasia grades at the intermediate time point with no dysplasia in the CBA/Ca + 4NQO model, severe dysplasia in the PIK3CA + 4NQO model and moderate dysplasia in the PIK3CA‐E6/E7 model. Ki‐67 expression at the ventral tongue sites also slightly increased between baseline and the intermediate time point in the transgenic models, while there was no apparent change in the CBA/Ca + 4NQO model.

FIGURE 6.

Immunohistochemical staining of dorsal tongue specimens from each transgenic model. Slight increases in Ki‐67 expression were observed between baseline and later time points in the PIK3CA + 4NQO and PIK3CA‐E6/E7 models. Scale bars represent 100 μm

4. DISCUSSION

This pilot study compares temporal changes in a wide range of biomarkers for the PIK3CA + 4NQO and PIK3CA‐E6/E7 experimental transgenic mouse models of oral carcinogenesis and an existing CBA/Ca + 4NQO model. Our results show that both transgenic models histologically resemble human OSCC and exhibit slightly increased expression of the cellular proliferation marker Ki‐67 at dysplastic sites. However, the PIK3CA + 4NQO model exhibits the most consistent lymphocytic infiltration, has prominent plaques and tumours on the dorsal and ventral tongue surface, and develops cancer more rapidly than alternative models. Additionally, as this model was developed to incorporate both genetic mutations and chemical carcinogens, it reflects the multiple factors which influence tobacco‐related human tongue cancer development.

Severe dysplasia in the PIK3CA + 4NQO model was first identified in the 3‐ to 8‐week time frame, earlier than it was observed in the existing CBA/Ca + 4NQO model (9‐12 weeks). Furthermore, the first instance of OSCC in the PIK3CA + 4NQO model occurred approximately 2 months earlier than in the CBA/Ca + 4NQO model. The PIK3CA‐E6/E7 model also showed severe dysplasia in the 3‐ to 8‐week time frame; however, OSCC was not observed by the end of the 20‐week time course for this model. While it is possible that the PIK3CA‐E6/E7 model may have eventually developed OSCC in a longer time course, all remaining mice from both transgenic models were sacrificed at 20 weeks to comply with IACUC humane endpoint guidelines. Altogether, the PIK3CA + 4NQO model's significantly shorter time course to OSCC reduces the time and resource burden associated with studies of oral carcinogenesis in animal models, allowing for a greater number of studies to be conducted.

In human oral cancer, intraoral carcinoma is commonly found on the ventral tongue, making ventral tongue dysplasia and cancer desired features in an animal model. ³¹ When ventral tongue histology was compared between models at the same time point, dysplasia was only observed in the transgenic models and not in the CBA/Ca + 4NQO model. Of all models, the PIK3CA + 4NQO model was the only to develop prominent ventral tongue tumours. The presence of dysplasia and OSCC in the ventral tongue distinguishes the PIK3CA + 4NQO model in its similarity to the gross visual characteristics of human disease. Furthermore, the development of prominent tumours in the PIK3CA + 4NQO model increases its potential utility in applications specifically targeting tumours, such as the intra‐tumoral injection of anticancer agents. ³²

Lymphocytic infiltration was most frequently and consistently observed in the PIK3CA + 4NQO model across sites with mild dysplasia, moderate dysplasia and severe dysplasia. We hypothesize that the combination of chemical carcinogen exposure and genetic mutations was the reason for the higher incidence of lymphocytic infiltration. Lymphocytic infiltration in the PIK3CA + 4NQO model suggests the model may be used to explore the role of immunosurveillance in oral carcinogenesis and for immunotherapy studies. ³³

The slight increase in Ki‐67 expression observed in the transgenic models between baseline and intermediate time points was consistent with the literature, as Ki‐67 is a cellular proliferation marker and has been reported to increase throughout the process of carcinogenesis. ³⁴ In this study, the limited number of specimens processed for immunohistochemical staining did not allow for a more comprehensive comparison across multiple time frames or dysplasia grades. Future immunohistochemical studies should be conducted to assess changes in the expression of other cancer‐related molecular biomarkers throughout the process of carcinogenesis. Additional work could also include staining for molecular biomarkers earlier in the time course, where expression may provide insights about the progression risk of sites with mild dysplasia.

While this was a small pilot study, it demonstrated that the PIK3CA + 4NQO model is a promising clinically translatable mouse model which closely mimics the morphological and molecular changes in human oral tongue carcinogenesis. Rapid cancer development is induced by a combination of chemical carcinogen and genetic influences and mimics the cytological and architectural features of human OSCC. Development of prominent changes in gross visual appearance facilitates research applications focused on OPL identification and tumour margins. The presence of lymphocytic infiltration suggests immune system involvement and presents opportunities for cancer immunology research applications. Positive staining of a candidate biomarker suggests that changes in molecular biomarker expression might be an area of interest for exploration. In sum, the PIK3CA + 4NQO model reflects the genetic, histologic, immunologic, molecular and gross visual features of human OSCC, making it a useful tool for future OSCC studies.

CONFLICTS OF INTEREST

The authors report no conflicts of interest.

Supporting information

Fig S1

Tan MT, Wu JG, Callejas‐Valera JL, et al. A PIK3CA transgenic mouse model with chemical carcinogen exposure mimics human oral tongue tumorigenesis. Int J Exp Path. 2020;101:45–54. 10.1111/iep.12347