Preclinical murine tumor models: a structural and functional perspective

Abstract

The goal of this review is to pinpoint the specific features, including the weaknesses, of various tumor models, and to discuss the reasons why treatments that are efficient in murine tumor models often do not work in clinics. In a detailed comparison of transplanted and spontaneous tumor models, we focus on structure-function relationships in the tumor microenvironment. For instance, the architecture of the vascular tree, which depends on whether tumor cells have gone through epithelial-mesenchymal transition, is determinant for the extension of the spontaneous necrosis, and for the intratumoral localization of the immune infiltrate. Another key point is the model-dependent abundance of TGFβ in the tumor, which controls the variable susceptibility of different tumor models to treatments. Grounded in a historical perspective, this review provides a rationale for checking factors that will be key for the transition between preclinical murine models and clinical applications.

Keywords: EMT; TGFβ; cancer biology; immunology; inflammation; microenvironment; spontaneous tumors; transplanted tumors; vascularization.

Figure 1.. Fibronectin expression is restricted to stromal cells in PyMT tumors, but not in tumors that have undergone EMT.

In the MMTV-PyMT model, fibronectin is only expressed by stromal cells, if the tumor is (A) spontaneous (SP) or (B) results from an orthotopical implant (TP) of cells dissociated from a SP tumor. Fibronectin may also be expressed by all tumor cells after EMT has taken place in (C) the TC1 cell line, here transplanted subcutaneously, or in (D) a spontaneous melanoma, TiRP. Tumor pieces were fixed overnight with periodate-lysine-paraformaldehyde at 4°C. Then, tumor samples were embedded in 5% low-gelling temperature agarose prepared in phosphate-buffered saline (PBS). 350 μm slices were cut with a vibratome in a bath of ice-cold PBS. Immunostaining of surface markers was performed at 37°C for 15 min with antibodies specific to EpCAM-BV421, CD31-Biot (both BD Pharmingen) or Fibronectin (Abcam). CD31-Biot was revealed with a streptavidin-PE (BD Pharmingen) and fibronectin with a goat anti-rabbit-FITC (Invitrogen).

Figure 2.. Growth rate of TP and SP PyMT tumors.

(A) The growth of TP tumors slows down with time. (B) The growth of SP tumors, which is initially slow, accelerates with time. Data from 45 tumors from 45 TP mice and 26 tumors from 9 SP mice. The small and large tumor diameters, d and D, were measured with calipers. Tumor volume was approximated by the volume of a spheroid (d².D/2). For each tumor, the volume at time day x is expressed relative to its volume at day 0, when the tumor volume was ~ 50–150 mm³. The ratios (volume(dx)/volume(d0)) expressed on a Log2 scale, are represented as a function of time.

Figure 3.. DMXAA induces a transient regression of TP melanomas but not of isogenic SP tumors.

(A) B10.D2 mice were injected subcutaneously with 2 million T429.11 cells, a cell line derived from an induced Amela TiRP tumor (Zhu et al., 2017) (TP tumors). (B) TiRP mice were treated subcutaneously with 2 mg of 4-OH tamoxifen to induce tumor development as described in Zhu et al. (2017) (SP tumors). Tumor growth was measured with calipers and approximated to the volume of a spheroid (d².D/2). When the average diameter was between 7 and 9 mm, T429-tumor-bearing mice or TiRP-tumor-bearing mice were randomized into two groups, with an average tumor volume of 220 mm³, and were treated with a single intraperitoneal (i.p.) injection of DMXAA (23 mg/kg) or DMSO (day 0). Tumor growth was monitored. Graphs show means ± S.E.M.