Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2015 Jun 17;6:7419.
doi: 10.1038/ncomms8419.

Characterization of Twenty-Five Ovarian Tumour Cell Lines That Phenocopy Primary Tumours

Affiliations
Free PMC article

Characterization of Twenty-Five Ovarian Tumour Cell Lines That Phenocopy Primary Tumours

Tan A Ince et al. Nat Commun. .
Free PMC article

Abstract

Currently available human tumour cell line panels consist of a small number of lines in each lineage that generally fail to retain the phenotype of the original patient tumour. Here we develop a cell culture medium that enables us to routinely establish cell lines from diverse subtypes of human ovarian cancers with >95% efficiency. Importantly, the 25 new ovarian tumour cell lines described here retain the genomic landscape, histopathology and molecular features of the original tumours. Furthermore, the molecular profile and drug response of these cell lines correlate with distinct groups of primary tumours with different outcomes. Thus, tumour cell lines derived using this methodology represent a significantly improved platform to study human tumour pathophysiology and response to therapy.

Conflict of interest statement

T.A.I. discloses pending patent intellectual proprietary interest as the inventor of OCMI media. The remaining authors declare no competing financial interests.

Figures

Figure 1
Figure 1. Long-term culture of ovarian tumours in OCMI media.
(a) Cells from malignant ascites fluid-containing ovarian carcinoma cells (OCI-U1a) were plated in OCMI (blue line), MCDB105/M199 with 10% serum (green line) or RPMI-1640 with 10% serum (red line). A significant decrease in population doubling is observed in MCDB105/M199 and RPMI-1640 media after 7 weeks (green and red lines). This was not observed in OCMI, in which the cells reached 30 population doublings at cessation of the experiment. A fast proliferating cell population emerged in RPMI-1640 medium around day 90 (red line) and was established as a permanent cell line. (b) The whole-genome DNA copy-number variants (CNVs) of the OCI-U1a tumour cells cultured in OCMI and RPMI and from the paired uncultured tissue were examined using a 250-K SNP array. The CNV trace reveals several peaks that are gained (amplifications, black arrows) and lost (deletions, grey arrows) in tumour cells cultured in RPMI (top trace) compared with the uncultured tumour sample (middle trace). In contrast, the CNV pattern of tumour cells cultured in OCMI appears significantly more similar to the uncultured tumour (bottom and middle CNV traces, respectively). (c,d) The ovarian clear-cell lines OCI-C4p and OCI-C5x were cultured to at least 60 and 100 population doublings, respectively, in OCMI (blue line). In the control medium (DMEM:F12 with 10% serum) the cells stopped proliferating in 30–40 days (red line). (e) Total cell numbers were plotted instead of population doublings to highlight the scale difference in OCMI (blue line) versus DMEM:F12 with 10% serum (red line) depicted in d. See Supplementary Methods, Supplementary Figs 1–6 and Supplementary Tables 1–7 for more detailed cell line and tumour information.
Figure 2
Figure 2. mRNA profiling of OCI cell lines identifies two major clusters.
(a) Unsupervised hierarchical clustering of mRNA expression levels of OCI (blue bars) and SOC (red bars) ovarian cancer cell lines reveal two clusters. Genes with ≥2-fold difference relative to the median value in at least four cell lines were selected for hierarchical clustering analysis (5,146 gene features). Rows represent individual genes and columns represent each cell line. The red and green colours reflect relative high and low expression levels, respectively, as indicated in the scale bar (log2-transformed scale). Two major clusters are observed: cluster 1 contains only OCI cell lines (left cluster, blue only), and cluster 2 contains a mixture of SOC and OCI cell lines (right cluster, red and blue). While the papillary serous histotype almost exclusively align within cluster 1 (green bars), the other subtypes are present in both clusters (orange bars). See Supplementary Methods, Supplementary Figs 10 and 11 and Supplementary Tables 8–11 for more detailed information. The complete mRNA expression data set is available at Gene Expression Omnibus, GEO accession number GSE40788. (b) The dendogram of the cell lines that make up the two clusters in the heatmap in a. The cell line names are coloured as follows; first column, OCI (blue) and SOC (red); second column, papillary serous (dark green) and other histotypes (orange); third column, papillary serous (dark green), clear cell (light blue), endometrioid (pink), mucinous (light green) and other histotypes (orange).
Figure 3
Figure 3. The proteomic profiling of OCI cell lines identifies two major clusters.
(a) The unsupervised clustering of protein expression (measured by RPPA) in OCI cell lines (blue bars) together with SOC ovarian cancer cell lines (red bars) reveal two distinct clusters. Rows represent cell lines and columns represent antibody probes for each protein. The red and green colours reflect relative high and low expression levels, respectively. As in the mRNA clustering, cluster 1 contains only OCI cell lines (top half of the heatmap, blue only), and cluster 2 contains a mixture of SOC and OCI cell lines (bottom half of the heatmap, red and blue). While the papillary serous histotype almost exclusively aligns within cluster 1 (green bars), the non-papillary serous subtypes (orange bars) are divided between cluster 1 and cluster 2. (b) The dendogram of the cell lines that make up the two clusters in a. The cell line names are coloured as follows; papillary serous (green) and other histotypes (orange). See Supplementary Methods, Supplementary Figs 12–14, Supplementary Tables 12 and 13 and Supplementary Data 4 and 5 for more detailed information.
Figure 4
Figure 4. Histopathology of OCI xenografts recapitulate the original human tumour.
(ac) Haematoxylin and eosin (H&E)-stained sections of primary human tumours used to establish OCI-P8p (papillary serous), OCI-E1p (endometrioid) and OCI-C3x (clear cell) cell lines. (df) H&E-stained sections of xenografts tumours derived by injecting SOC cells (ES2, SKOV3 and TOV-112D) subcutaneously into immunocompromised mice. The typical features of human adenocarcinomas such as glands, papillae, stromal cores and desmoplastic stroma are absent. (go) H&E-stained sections of xenograft tumours derived by injecting OCI cell lines (P5x, P7a, P9a, C5x, C3x, CSp and E1p) subcutaneously into immunocopromised mice. In papillary serous specimens, note the presence of stromal cores and papillary architecture (gi). In the endometrioid specimen note the presence of glands (m) which were positive for oestrogen receptor (ER) and mucin (brown), respectively, consistent with the endometrioid phenotype (n,o). Scale bar, 100 μM. See Supplementary Fig. 14 for additional images.
Figure 5
Figure 5. Histotype-specific immunostains of SOC and OCI xenograft tumours.
Formalin-fixed paraffin-embedded sections of xenograft tumours derived from OCI and SOC lines were stained with ovarian carcinoma histotype-specific markers; PAX8, oestrogen receptor (ER), Wilm's tumour (WT1), p53, p16, HNF1β and ARID1a. Nuclear staining (brown) for PAX8, ER, WT1, p53, HNF1β and ARID1a is specific. The occasional cytoplasmic or stromal staining with these markers is nonspecific. In contrast, tumour-specific p16 staining is typically cytoplasmic. Scale bar, 20 μM. See Supplementary Fig. 15 for additional images.
Figure 6
Figure 6. The OCI lines identify two major outcome groups.
(a) Unsupervised hierarchical clustering of gene expression data of 37 cell lines and 285 human tissues reveal three clusters. Genes with ≥2-fold expression differences relative to the median value across cell lines and tissues in at least four samples were selected for hierarchical clustering analysis (3,831 gene features). Rows represent individual genes and columns represent each cell line or tissue. The red and green colour in cells reflect relative high and low expression levels, respectively, as indicated in the scale bar (log2-transformed scale). Red and blue bars above the heatmap indicate human tumour samples; light-blue bars indicate OCI lines; and black bars indicate SOC lines. While SOC cell lines (black bars) are exclusively group within patient cluster P2 (red bar), the OCI cells (light-blue bars) predominantly group within patient cluster P1 (blue bar). A small subset of tumour samples form a third cluster that does not include any cell lines (white); this group was excluded from outcome analysis. The complete mRNA expression data set is available at Gene Expression Omnibus, GEO accession number GSE40788 and GSE9899. (b) The progression-free and overall survival analysis data of patients with the ovarian tumours in patient cluster 1 and 2 in a. The patients with tumours that have a gene expression profile to OCI lines (blue bar, cluster 1 in a) have worse outcomes than patients with tumours that have gene expression profiles similar to SOC lines (red bar, cluster 2 in a). The small cluster of tumours that did not include any cell lines (white) was excluded from the outcome analysis.
Figure 7
Figure 7. The OCI lines identify two major taxol and cisplatin response groups.
(a) Taxol response of OCI and SOC cell lines that are in mRNA/RPPA cluster 1 versus cluster 2. The OCI and SOC lines were plated in triplicates in OCMI medium (3,000 cells per well) in 96-well plates. The next day, 20 nM taxol was added and metabolic activity was measured as 590/530 fluorescence via Alamar Blue after 96 h. OCI cell lines in mRNA/RPPA cluster 1, blue bars; SOC cell lines in cluster 1, red bars; OCI lines in cluster 2, white bars. The results are representative of four different experiments. See Supplementary Figs 18–20 for further details. (b) Proteins that are differentially expressed in mRNA/RPPA cluster 1. Analysis of RPPA data from OCI and SOC lines revealed a subset of proteins and phosphor-proteins that are differentially expressed in the taxol-resistant OCI lines (cluster 1, blue labels; cluster 2, red labels; P<0.05, Student's t-test). Rows represent cell lines and columns represent antibody probes for each protein. The red and green colours reflect relative high and low expression levels, respectively. Please see Supplementary Tables 14 and 15 and Supplementary Data 6 for further details.
Figure 8
Figure 8. The OCI cluster 1 is similar to TCGA and AOCS poor-outcome groups.
(a) Pearson correlation values for each of the signature scores indicate a very strong correlation among the genes that are expressed in the OCI cluster 1 and the Tothill C1 group (R=0.93). Interestingly, both of these signatures also exhibit a strong correlation with genes that are abundantly expressed in vascular endothelial cells (>0.7), suggesting that both of these signatures define ovarian cancer cells with mesenchymal/endothelial attributes. (b) The Kaplan–Meier overall survival plot of the Tothill C1–C6 groups indicate that the Tothill C1 group that correlated with OCI cluster 1, has one of the poorest outcomes. (c) The box-and-whisker plots of the signature scores for the TCGA ‘poor prognosis genes' demonstrate that the TCGA mesenchymal group is most similar to OCI cluster 1 in a statistically significant manner. The two-tailed t-test was used to compare the mesenchymal group signature score versus the three other groups (P=0.034). The interquartile range is shown by the box and the bar within the box represents the median value. (d) The Kaplan–Meier overall survival plot of the 482 TCGA tumours with clinical follow-up demonstrate that the TCGA mesencyhmal subtype that correlates with the OCI cluster 1 also has one of the worst outcomes. See Supplementary Figs 21 and 22 for further details.

Similar articles

See all similar articles

Cited by 54 articles

See all "Cited by" articles

References

    1. Gey G. O., Coffman W. D. & Kubicek M. T. Tissue culture studies of the proliferative capacity of cervical carcinoma and normal epithelium. Cancer Res. 12, 264 (1952).
    1. Voskoglou-Nomikos T., Pater J. L. & Seymour L. Clinical predictive value of the in vitro cell line, human xenograft, and mouse allograft preclinical cancer models. Clin. Cancer Res. 9, 4227–4239 (2003). - PubMed
    1. Shoemaker R. H. The NCI60 human tumour cell line anticancer drug screen. Nat. Rev. Cancer 6, 813–823 (2006). - PubMed
    1. Neve R. M. et al. A collection of breast cancer cell lines for the study of functionally distinct cancer subtypes. Cancer Cell 10, 515–527 (2006). - PMC - PubMed
    1. Verschraegen C. F. et al. Establishment and characterization of cancer cell cultures and xenografts derived from primary or metastatic Mullerian cancers. Clin. Cancer Res. 9, 845–852 (2003). - PubMed

Publication types

Associated data

Feedback