An in vivo tumour organoid model based on the chick embryonic chorioallantoic membrane mimics key characteristics of the patient tissue: a proof-of-concept study

Abstract

Background: Patient-derived tumour organoids (PDOs) are highly advanced in vitro models for disease modelling, yet they lack vascularisation. To overcome this shortcoming, organoids can be inoculated onto the chorioallantoic membrane (CAM); the highly vascularised, not innervated extraembryonic membrane of fertilised chicken eggs. Therefore, we aimed to (1) establish a CAM patient-derived xenograft (PDX) model based on PDOs generated from the liver metastasis of a colorectal cancer (CRC) patient and (2) to evaluate the translational pipeline (patient - in vitro PDOs - in vivo CAM-PDX) regarding morphology, histopathology, expression of C-X-C chemokine receptor type 4 (CXCR4), and radiotracer uptake patterns.

Results: The main liver metastasis of the CRC patient exhibited high 2-[¹⁸F]FDG uptake and moderate and focal [⁶⁸Ga]Ga-Pentixafor accumulation in the peripheral part of the metastasis. Inoculation of PDOs derived from this region onto the CAM resulted in large, highly viable, and extensively vascularised xenografts, as demonstrated immunohistochemically and confirmed by high 2-[¹⁸F]FDG uptake. The xenografts showed striking histomorphological similarity to the patient's liver metastasis. The moderate expression of CXCR4 was maintained in ovo and was concordant with the expression levels of the patient's sample and in vitro PDOs. Following in vitro re-culturing of CAM-PDXs, growth, and [⁶⁸Ga]Ga-Pentixafor uptake were unaltered compared to PDOs before transplantation onto the CAM. Although [⁶⁸Ga]Ga-Pentixafor was taken up into CAM-PDXs, the uptake in the baseline and blocking group were comparable and there was only a trend towards blocking.

Conclusions: We successfully established an in vivo CAM-PDX model based on CRC PDOs. The histomorphological features and target protein expression of the original patient's tissue were mirrored in the in vitro PDOs, and particularly in the in vivo CAM-PDXs. The [⁶⁸Ga]Ga-Pentixafor uptake patterns were comparable between in vitro, in ovo and clinical data and 2-[¹⁸F]FDG was avidly taken up in the patient's liver metastasis and CAM-PDXs. We thus propose the CAM-PDX model as an alternative in vivo model with promising translational value for CRC patients.

Keywords: 2-[18F]FDG; CAM; CRC; In ovo; PDX; PET/MRI; Patient-derived organoids; [68Ga]Ga-Pentixafor.

Fig. 1

Experimental workflow of the study. Created with BioRender.com

Fig. 2

Diagnostic abdominal 2-[¹⁸F]FDG and [⁶⁸Ga]Ga-Pentixafor PET/MR images of the liver metastasis in the CRC patient. (A, D) Contrast-enhanced T₁-VIBE MR images, (C, F) PET images and (B, E) fused PET and MR images are presented in axial views for 2-[¹⁸F]FDG (105 min p.i., 198 MBq, upper row) and [⁶⁸Ga]Ga-Pentixafor (40 min p. i., 158 MBq, lower row). Long blue arrows indicate the liver metastasis and short blue arrows indicate focal [⁶⁸Ga]Ga-Pentixafor uptake

Fig. 3

Histopathological characterisation of CAM-PDXs and comparison with the patient’s liver metastasis and in vitro PDOs. H&E, PAS stain, and IHC for CK AE1/AE3, Ki67, and CXCR4 are shown for CAM-PDXs (n ≥ 3), patient’s liver metastasis (n = 1), and PDOs (n = 1). Additional IHC for CAM-PDXs for Desmin and CC3 (n ≥ 4) is shown. Magnification: 200-fold; 40-fold (Desmin IHC (left)). The 200x Desmin image corresponds to the central part of the 40x magnification as indicated by a black rectangle

Fig. 4

PDOs in culture and [⁶⁸Ga]Ga-Pentixafor PDO uptake. Representative bright-field microscopic images of (A) pre-ovo and (B) post-ovo PDOs. Magnification: 40-fold; scale bars correspond to 500 μm. (C) [⁶⁸Ga]Ga-Pentixafor uptake into pre-ovo PDOs compared to post-ovo PDOs shown as a percentage of applied dose per 1.5 × 10⁵ cells (%AD/1.5 × 10⁵ cells). Significantly different datasets are marked with an asterisk (* p < 0.05, ** p < 0.01)

Fig. 5

Images of [⁶⁸Ga]Ga-Pentixafor and 2-[¹⁸F]FDG PET/MRI in the CAM-PDX model of CRC. (A) [⁶⁸Ga]Ga-Pentixafor baseline (EDD16; 4.1 MBq, 1.6 nmol [⁶⁸Ga]Ga-Pentixafor), (B) blocking (EDD17; 7.2 MBq, 2.3 nmol [⁶⁸Ga]Ga-Pentixafor; 315 µg AMD3100), and (C) 2-[¹⁸F]FDG scans (EDD18; 5.8 MBq) were performed using the same subject on three consecutive days. Left: T₂-TurboRARE MR images, right: PET images 60 min p.i., middle: fused PET/MR images in axial view. PET data are provided as a percentage of total activity per cm³ (%TA/cc). The xenograft is marked with yellow arrows