Can experimental models of rodent implantation glioma be improved? A study of pure and mixed glioma cell line tumours.

To evaluate the hypothesis that co-implantation of different rodent glioma cell lines might result in experimental brain tumours that more closely resemble human gliomas the neuropathology and immunocytochemical features of implantation gliomas derived from single cell lines (C6, A15A5, F98), two cell lines admixed 50:50 prior to implantation (C6 + F98 and C6 + A15A5) and three cell lines equally admixed (C6 + A15A5 + F98) was studied in the adult Wistar rat. Tumours grew consistently following implantation of the single and the two admixed cell lines, however tumour growth following triple mix implantation was considerably and consistently impaired. The tumours derived from admixed cell lines showed regional heterogeneity with areas characteristic of both the primary cell lines. Foci of lymphocytic infiltrates, tumoural necrosis, often with pseudopallisading, and peritumoural edema were consistent features of all tumours. Limited parenchymal and more extensive perivascular tumoural invasion was seen predominantly in tumours containing the C6 cell line. There were no significant differences in GFAP, vimentin and HSP70 staining between the mixed tumours, although the pure F98 and A15A5 tumours were, unlike the pure C6 gliomas, S-100 negative. Using PCNA expression as a measure of the tumour proliferation all except the tumours derived from the three cell lines mix, which had a staining index of 7-10%, had focal staining indices in viable tumour of between 40-80%. There was focal positive staining in both perilesional brain and in regions of all tumours for the macrophage markers ED-1 and ED-2. None of the three cell lines stained in vitro for either ED1 and ED2 but all were constitutively positive in vitro for OX-6, a proposed marker for antigen presenting cells. The macrophage and lymphocytic response suggest a vigorous but largely ineffective immunological response had been mounted against all tumours. The consistent failure of the triple mix tumours to grow is unexplained. This work has shown the feasibility of producing 'mixed' cell line experimental gliomas by combining two cell lines at the time of innoculation. However, the relative failure to produce (i) mixed tumours that have properties not inherent to either parent cell line and (ii) implantation glioma with three cell lines suggest there are limits to this approach. Admixture of cell lines at the time of implantation therefore does not make experimental glioma models that more closely resemble natural gliomas, and also has some particular disadvantages. This experimental approach is therefore not recommended for use in the study of tumour biology and in evaluating the effectiveness of novel therapies.