Three-dimensional in vitro cell biology models of ovarian and endometrial cancer

Abstract

Objectives: This study aims to establish three-dimensional (3D) cell culture models of human ovarian and endometrial cancers and to compare biological and morphological characteristics of these models with those of two-dimensional (2D) models of the same cell lines and the primary tumours.

Methods: 3D models of ovarian and endometrial cancer cell cultures were established using a Rotary Cell Culture System. Immunohistochemical profiling and differential proteomics were used to characterize biological characteristics of multicellular spheroids (MCS) formed from these cultures. These were compared to characteristics of the same cells established in 2D and of the primary tumours from which the cell lines were derived.

Results: MCSs from 3D cell cultures appeared histologically similar to the primary tumours. Immunohistochemical profiling of multiple markers, including CA125, BCL2 and p53, showed that patterns of protein expression in MCSs resemble those of the primary tumours. Proteomic profiling identified several differentially expressed protein markers between 2D and 3D cultures. These included prohibitin, which was down-regulated in 3D cultures suggesting cells proliferate less compared to 2D cultures; and VDAC1 and annexin 4, which were up-regulated in 3D cultures suggesting greater levels of apoptosis in 3D compared to 2D models.

Conclusion: Establishing 3D models of cancer cell lines is likely to be of value for studying the molecular and biological mechanisms of ovarian/endometrial tumour progression and for testing novel molecular targets for cancer therapy.

Figure 1

Growth, neoplastic and epithelial characteristics of the cell lines OV‐TRL12B and EN‐TRL67T grown in two‐dimensional (2D) cultures. (a) Growth curves of OV‐TRL12B () and EN‐TRL67T () over a period of 60 days in culture. (b) Colony‐forming efficiency (CFE%) and (c) invasive ability (cell number) of OV‐TRL12B and EN‐TRL67T cell lines. OV‐TRL12B cells showed a 3% CFE, EN‐TRL67T a CFE of 0.5%, and both cell lines also showed evidence of invasive ability.

Figure 2

(a–c) Three‐dimensional (3D) cultures of the OV‐TRL12B cancer cell line (a and b) show cell clusters composed of sheets of cells showing variation in cell size. At higher magnification, nuclear pleomorphism and prominent nucleoli (N) are seen. Histological section of the original tumour (c) shows poorly differentiated squamous carcinoma infiltrating stroma, forming sheets and nests of cells. Tumour cells show similar cytological features with variation in cell size nuclear pleomorphism and prominent nucleoli, as the 3D culture; (d–f) 3D cultures of the EN‐TRL 67T cancer cell line (d and f) shows papillary formation of the cell clusters. Tumour cells show nuclear pleomorphism, increased mitotic activity and atypical mitoses (M). Foci of apoptosis (A) and nuclear debris are present. Histological section of the original endometrial adenocarcinoma (f) shows papillary architecture and prominent neutrophil polymorph (P) infiltration of stroma. Tumour cells show similar architectural and morphological features although apoptosis is more prominent in the 3D culture.

Figure 3

(a) Immunohistochemical analysis of OV‐TRL12B and EN‐TRL67T cell lines grown in two‐ (2D) and three‐dimensional (3D) cultures and the corresponding primary tumour staining for the cytokeratin marker CK7, the epithelial antigen BerEp4 and the epithelial tumour marker Ca125. (b) Summary of immunohistochemical staining for several markers. Shading of boxes is graded according to percentage of positive stained cells and staining intensity.

Figure 4

(a) Two‐dimensional difference gel electrophoresis (2D‐DIGE) analysis of the cancer cell lines OV‐TRL12B and EN‐TRL67T grown in two‐ (2D) and three‐dimensional (3D) cultures. The further validated proteins are marked and a magnification of these spots is shown. (b) Three proteins identified as being differentially expressed between 2D and 3D cultures of OV‐TRL12B and EN‐TRL67T cancer cell lines by 2D‐DIGE, validated by Western blotting analysis.