We have read with great interest the article by Shinagawa et al.1 In their study, they evaluated transplantation of the KM12SM human colon cancer cell line mixed with mesenchymal stem cells (MSCs) into the spleen.They reported a significantly greater number of liver metastases at 4 weeks with the cell mixture than with the transplantation of KM12SM cells alone. Additionally, MSCs surrounding the tumour were functionally incorporated into the stroma of the orthotopic colon tumour and expressed carcinoma-associated fibroblast markers. This manuscript prompted us to analyze the relevance of the murine model. To induce liver metastasis, a human colon cancer line (LS174) was injected into the spleen of nude mice. First, we characterized the bone marrow-derived cell (BMDC) population by fluorescence-activated cell sorting isolated on green mice C57 BL/6-Tg obtained after a density gradient. The population contained 30% of BM-derived endothelial precursor cells and derived haematopoietic precursor cells that expressed CD 133. At day 0, we injected BMDCs or PBS into the tail veins of tumor-bearing mice. At day 38, the injection of BMDCs enhanced the metastatic process in our model of liver metastasis with a number of micro or macrometastases in 56% of the BMDC group (n = 16) versus 20% in the PBS group (n = 20), p < 0.05. Immunofluorescent staining of cut liver specimens has demonstrated that BMDCs labelled with Green fluorescent protein (GFP) always localized to the tumour invasion front and colocalized with endothelial cells (Fig. 1). Migration of BMDCs to the stroma of a metastatic liver tumor (*): GFP-labeled BMDCs are localized to the tumor invasion front (a) and colocalized with endothelial cells (CD31 labeling in red) (b). (c) Merge (green and red). There is increasing evidence that BMDCs participate in the tumour microenvironment and metastatic progression,2-5 possibly by previous modification of the metastatic site by the BM-derived precursor cell before tumour cell arrival,4 which has led to the emergence of the concept of the premetastatic niche. We have corroborated the findings of Shinagawa et al. with a second type of human colon cancer cell line (LS174) in another original murine model of orthotopic liver metastasis. This method of BMDC supply was even systemic, which is more interesting than coinjection because the BMDCs localize preferentially and not around the tumour by necessity. Our choice of BMDCs was innovative due to the absence of sorting of the different cells that constitute the injected solution; this solution was at least a third bone marrow-derived endothelial and haematopoietic precursors cells. For the first time, we have proposed a functional role for bone marrow-derived endothelial and haematopoietic precursors cells in tumour angiogenesis and metastasis enhancement in an orthotopic nude mice model. Controlling the effects of BMDCs and understanding the molecular pathways of interaction between tumour cells and BMDCs or MSCs could be a new targeted therapy for liver colon metastasis. Yours sincerely, Raphaelle Audollent*, Clarisse Eveno* , Jean-Olivier Contreres*, Patricia Hainaud*, Aurore Rampanou*, Evelyne Dupuy* , Jean-Philippe Brouland* §, Marc Pocard* , * Unité Inserm U965, équipe “Angiogenèse et Recherche Translationnelle,” hôpital Lariboisière, Institut des Vaisseaux et du Sang, 8, rue Guy-Patin, 75475 Paris cedex 10, France, Département médico-chirurgical de pathologie digestive, Hôpital Lariboisière, 2, rue Ambroise-Paré, 75475 Paris cedex 10, France, Université Paris-Diderot, Paris-7, France, § Département d'Anatomopathologie, Hôpital Lariboisière, 2, rue Ambroise-Paré, 75475 Paris cedex 10, France.