Angiogenesis in patient‐derived xenografts of odontogenic myxoma

Abstract

Previously, by employing 3D organotypic tissue culture and patient‐derived xenograft (PDX) model, oral myxoma response to a MAPK/MEK inhibitor was observed. Gross examination of the tumour fragments obtained after 55 days of PDX grafting revealed increased capsule vascularization. Microscopic analyses showed blood capillaries intermixed with myxoma cells, but the origin of these capillaries, from mice or humans, was not established. This study aimed to investigate whether the endothelial cells observed in the myxoma PDX model are derived from the mouse or from the primary human tumour. Immunohistochemistry was performed on five tumour fragments from the PDX of myxoma after 55 days of implantation in mice. Immunopositivity for antibodies against human (HLA‐ABC) and mouse (H2 Db/H2‐D1) major histocompatibility complex class I (MHCI) was assessed in the endothelial cells. The endothelial cells in the PDX fragments revealed a membrane staining for the human MHCI protein in the PDX tumour and adjacent connective tissue capsule, indicating that capillaries were derived from the human tumour fragment. Considering the probable human origin of the endothelial cells from capillary blood vessels in the myxoma PDX, we conclude that this PDX model is an interesting model to study myxoma angiogenesis.

1. INTRODUCTION

Odontogenic myxoma (OM) is a benign odontogenic tumour microscopically characterized by stellate and spindle‐shaped cells dispersed in an abundant myxoid extracellular matrix (MEC). ¹ Despite its benign nature, the OM shows local invasiveness and aggressive growth, which often can be associated with tooth displacement, bone and dental resorption, and cortical expansion. ¹ , ² It mainly affects young individuals between the second and fourth decades of life. In addition, conservative surgical treatment is associated with higher recurrence rates. ² The molecular pathogenesis of OM remains unclear, despite the efforts of several research groups. ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸

Our group has recently reported MAPK/ERK pathway activation in OM patient samples. ⁹ In addition, a three‐dimensional organotypic tissue culture and a patient‐derived xenograft (PDX) model of OM were established and treated with U0126, a MEK1/2 inhibitor, which resulted in decreased cell viability and tumour shrinkage. On gross examination, tumour fragments removed after 55 days of PDX xenotransplantation establishment showed increased fibrous capsule vascularization. Notably, microscopic analysis revealed the presence of blood capillaries intermixed with OM cells. ⁹ These findings raised the question of whether these capillaries would be derived from the mouse or from the implanted primary human tumour. Here, we assessed the probable origin of endothelial cells observed in the capillary vessels of the PDX model.

2. MATERIALS AND METHODS

2.1. Tissue sampling and ethical approval

A fresh OM sample derived from a 29‐year‐old male patient was taken from the patient's left mandibular body, after he had consented to participate in the study and signed an informed consent form. All procedures were done in accordance with the Helsinki Declaration (1964) and its later amendments. The study was approved by the University of Minas Gerais Ethics Committee (CAAE 48 121 415.2.0000.5149). All procedures involving animals were performed in accordance with international guidelines and Brazilian regulations and were approved by the Institutional Animal Care and Use Committee (201/2015) of Universidade Federal de Minas Gerais.

2.2. Patient‐derived xenograft

The minimally processed tumour fragments of OM (n = 5) were obtained and subcutaneously implanted in mice to establish the patient‐derived xenograft (PDX) model and obtained as described previously. ⁹ , ¹⁰ The fragments were collected 55 days after implantation into three 8‐week‐old male NUDE BALB/c mice. On this day, the size of the PDX tumour fragments was significantly smaller than at implantation. Two of these mice were implanted with two tumour fragments each, while one mouse was implanted with a single minimally processed tumour fragment. ⁹ After removal, tumours were formalin‐fixed and paraffin‐embedded (FFPE).

2.3. Immunohistochemistry analysis

We used Immunohistochemistry (IHC) to target mouse and human MHC class I protein to elucidate the human or murine origin of the marked cells, specifically focusing on the endothelial cells. The FFPE samples were cut into 4‐μm‐thick sections. The IHC reactions followed standard procedures. Antigen retrieval was performed with Trilogy™ reagent at 1:100 (Cell Marque, Germany). Tissue sections were incubated with mouse monoclonal antibody targeting murine MHC Class I (H2 Db/H2‐D1; 28‐14‐8 clone, Abcam, UK) and human MHC Class I (HLA‐ABC; EMR8‐5 clone, Abcam, UK), diluted at 1:50 and 1:400 respectively. After 2 h of incubation at room temperature in a humid chamber, the reaction was visualized using the HRP‐labelled polymer‐based system EnVision+ (Dako Corporation, Carpinteria, USA) with diaminobenzidine (Sigma, St Louis, USA) as the chromogen. The slides were counterstained with Mayer's haematoxylin. Primary antibodies were omitted in negative controls. Samples with known immuno‐expression for the screened antigens (mouse spleen and human oral fibrous hyperplasia samples) served as positive controls.

3. RESULTS

The samples revealed membrane staining for the anti‐HLA‐ABC antigen in the endothelial cells of capillary vessels, suggesting that these capillaries are derived from the human tumour fragment implanted in the PDX (Figure 1A and Figure 1B). There was an absence of immunoexpression for the anti‐H2 Db/H2‐D1 antigen (Figure 1C and Figure 1D) in the endothelial cells of capillary vessels. Positive and negative controls representative images are shown in Figure 1E‐H. Analysis of the connective tissue capsule surrounding the tumour fragment after the PDX establishment revealed the presence of capillary vessels with strong staining for anti‐HLA‐ABC antibody and negative immuno‐expression for the anti‐H2 Db/H2‐D1 antibody (Figure 2).

FIGURE 1.

Immunohistochemical reactivity for HLA‐ABC and H2 Db/H2‐D1 in odontogenic myxoma PDX. Immuno‐expression for human MHCI (HLA‐ABC) in blood capillaries is shown in (A and B). The absence of immuno‐expression for mouse MHCI (H2 Db/H2‐D1) is shown in (C and D). Human oral fibrous hyperplasia was used as positive control for the human MHCI (E) and served as negative control for the murine MHCI expression (F). Mouse spleen samples showed no reactivity for the human MHCI protein (G) and served as positive control for the murine MHCI expression (H). Original magnifications a, c, e, f, g, h (10X) and b, d (20×)

FIGURE 2.

Immunohistochemical reactivity for HLA‐ABC and H2 Db/H2‐D1 in the connective tissue capsule surrounding the odontogenic myxoma PDX. Immuno‐expression for human MHCI (HLA‐ABC) in blood capillaries is shown in (A). The absence of immuno‐expression for murine MHCI (H2 Db/H2‐D1) is shown in (B). Original magnifications: 10×

4. DISCUSSION

The capacity of tumour cells to induce the sprouting of new vessels from pre‐existing ones, known as angiogenesis, constitutes one of the better established hallmarks of cancer. ¹ It is well known that solid tumours are capable of inducing angiogenesis, and tumour growth is itself dependent on this angiogenesis. ¹¹ , ¹² In a process called angiogenic switch, tumour cells foster the induction of new capillaries to provide nutrients, oxygen and the clearance of metabolic wastes, sustaining tumour growth. ¹¹ When establishing a PDX model for OM, we observed an abundance of capillary vessels, both permeating the tumour cells and the fibrous capsule that formed surrounding the tumour fragments. ⁹ Given the benign nature of OM, we first hypothesized that these capillaries most probably originated from the mice in which the PDX had been transplanted. After assessing by IHC the expression of mouse and human MHCI, where we observed the expression of the human protein in the endothelial cells of the tumour and fibrous capsule capillaries, we believe that the source of the cells was the transplanted human tumour fragments.

The immuno‐expression and role of several proteins involved in angiogenesis have been assessed in OM, encompassing: vascular endothelial growth factor A (VEGFA) family, orosomucoid‐1 protein (ORM1), metalloproteinases (MMPs), fibroblast‐derived growth factor 8 (FGF8), bone morphogenetic protein (BMP4) and proteins from the MAPK signalling pathway. ⁵ , ⁹ , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ The expression of MMP2 and MMP9 in OM samples was detected in vivo and in vitro. A zymographic assay demonstrated the presence of MMP9 in both active and latent forms, suggesting this enzyme is involved not only in the invasive growth of OM but also in angiogenesis. ¹³

High immuno‐expression of VEGFA and ORM1 by OM tumoral and endothelial cells was observed in 33 cases of the tumour, which are both potential contributors to tumour growth. ¹⁵ Furthermore, immuno‐expression patterns of FGF8 and BMP4 in OMhave been investigated, revealing nuclear positivity for FGF8 and BMP4 in 90% (9/10) and 80% (8/10) of the tumours respectively. ¹⁶ In addition, the phosphorylated active form of ERK1/2 protein, pERK1/2, was reported in both neoplastic and endothelial cells in 9 of 9 OM cases, consistent with MAPK/ERK pathway activation in these lesions. ⁹ According to a computational model constructed, the combined effects of FGF and VEGF are related to ERK phosphorylation, leading to MAPK/ERK pathway activation. ¹⁸ As no mutation in gene components of MAPK/ERK pathway was detected in OM, ⁷ MAPK/ERK pathway activation may be related to VEGF and FGF.

Malignant tumours have well‐established capability to co‐opt and incorporate the host tissue microvasculature to support their growth. ¹¹ , ¹⁹ Interestingly, in the OM PDX model we established the immunohistochemical results suggest that the blood capillaries in the transplanted tumour and connective tissue capsule are derived from the patient tissue, not from the mice. Thus, we hypothesize that the limited potential of OM cells to recruit host endothelial cells to form a vascular network may contribute to its benign nature. Conversely, the capability of OM cells to migrate through the connective tissue capsule and recruit human mesenchymal / endothelial cells from the tumour fragments implanted in the PDX model could be related to its aggressive and locally invasive clinical behaviour.

Vasculogenic mimicry describes the process whereby cancer cells become capable of generating vascular‐like structures, regardless of the presence of endothelial cells. ²⁰ , ²¹ , ²² In line with this process, many human capillaries were observed between tumour cells and in the surrounding fibrous capsule. Many of the signalling pathways and molecules involved in this process have been shown to be shared with those related to OM pathogenesis and progression (e.g. MMPs, VEGFA and MAPK‐ERK signalling), and are often related to hypoxia. ²⁰ , ²³ , ²⁴ To the best of our knowledge, there is no evidence of vascular mimicry occurring in benign neoplasms. However, given the locally invasive, destructive growth pattern and high recurrence rates of OM, ² , ²⁵ the plausibility of vascular mimicry in OM should be further investigated.

Notably, it is important to discuss the contribution of the mouse microenvironment to angiogenesis. A set of bone marrow cells, including those from the innate immune system, play roles in angiogenesis in pathological processes. ¹¹ , ¹² , ²⁶ Despite the partial absence of the adaptive immune system, NUDE mice have an innate immune system, harbouring natural killer (NK) cells, granulocytes, dendritic cells and B cells. ²⁷ , ²⁸ , ²⁹ , ³⁰ It is possible that in NUDE mice innate immune cells contributed to angiogenesis, suggesting an interplay between innate immune cells of NUDE mice and the pro‐angiogenic factors secreted by OM cells.

Thus we hope that this pilot study and our preliminary results will instigate new studies to elucidate the mechanisms of tumour vascularization in OM.

5. CONCLUSION

Considering the probable human origin of the endothelial cells from capillary blood vessels in the OM PDX model, we conclude that this model is interesting and suitable in order to study OM tumour angiogenesis.

FUNDING INFORMATION

The authors are thankful for the support of the Research Support Foundation of the State of Minas Gerais (FAPEMIG, Brazil), the National Council for Scientific and Technological Development (CNPq, Brazil) and the Coordination for the Improvement of Higher Education Personnel (CAPES, Brazil). VCB receives a CAPES scholarship, and JCS receives a CNPq scholarship. RSG and CCG are research fellows at CNPq.

CONFLICT OF INTEREST

The authors declare that there are no conflicts of interest.

Souza JCD, Bastos VC, Pereira NB, et al. Angiogenesis in patient‐derived xenografts of odontogenic myxoma. Int J Exp Path. 2022;103:65–69. doi: 10.1111/iep.12431