Effects of human bone marrow stroma on the growth of human tumor cells.

As some tumors metastasize frequently to marrow we modified the clonogenic assay for human tumor cell growth by culturing tumor cells in the presence of human bone marrow stromal cells. In a bilayer soft agar assay, human tumor cells which had been passaged in nude mice were plated in the agar overlayer on an underlayer containing a suspension of trypsinized human bone marrow stromal cells. These marrow stromal cells stimulated the growth of tumor cells in a dose-dependent fashion, with a growth peak at a stromal cell density of 5-10 x 10(5)/ml. The maximal stimulation of tumour cell growth was 13-fold. We evaluated clonal growth of six separate tumors of five different histological types (small and large cell bronchogenic carcinoma; mammary carcinoma; malignant melanoma; pleural mesothelioma) and demonstrated that in 9 of 11 experiments tumor cell colonies formed in the absence of stromal cells, but colony growth was markedly stimulated by stromal cells in every case. Stromal stimulation persisted after irradiation of the stromal cells with 10 Gy. Growth of five fresh human tumor samples was similarly stimulated by the presence of human bone marrow stromal cells. Tumor cell colonies were characterized morphologically by Pappenheim stain and immunologically for surface antigens by peroxidase-antiperoxidase immunostaining utilizing monoclonal antibodies (carcinoembryonic antigen 26/3/13 and 26/5/1, EMA, HEA125, Sam 2 and Sam 10) which detected epithelial cell antigens. Colonies consisted of cytologically malignant cells which expressed epithelial cell antigens. Thus, the tumor cell origin of colonies from mammary carcinoma and bronchogenic small cell, large cell, and adenocarcinoma was proven. This tumor stem cell assay permits further analyses of human tumor cell biology and may be useful for testing drug sensitivity.