Modeling and histopathological recognition of anoikis resistance in colorectal carcinoma.

Colorectal cancer (CRC) is the third most common cause of cancer related deaths in Finland and shows increasing incidence. Anoikis is a special form of cell death. In normal cells, loss of contact with the extracellular matrix induces anoikis. Ability to resist anoikis is a common abnormality in cancer cells and one of the most important mechanisms by which cancer cells can survive after detaching from their original site, and therefore essential in the formation of metastases. The importance of anoikis resistance in cancers is known for years. It can only be detected with cell culture based tests necessitating specimens with living cells and been studied in few human cancers. Lack of reports on detection of anoikis resistance, determination of its severity in assessing prognosis and selecting optimal treatment for patients is alarming. Furthermore, the biological mechanism behind anoikis resistance remains largely unknown.

Our study was based on the hypothesis that in actual human cancers anoikis resistance allows carcinoma cells to form multicellular clusters, where the inner cells loss their contact with the extracellular matrix, but still survive. Detection of such clusters in conventional pathological specimens of carcinomas could indicate that such tumors shows anoikis resistance. We tested the hypothesis by cell culture experiments with different colon cancer cell lines. In addition, we tested the detection such clusters in conventional specimens of colorectal carcinoma cases and evaluated prognostic significance of abundance and types of such clusters.

We found that Caco-2 colon carcinoma cells did not survive in suspension cultures indicating that they are not resistant to anoikis. In 3-D culture these cells formed cysts with consistent absence of multicellular clusters, this growth patterns being consistent with the lack of anoikis resistance. In contrast, transfection with mutated KRAS or BRAF oncogenes to Caco-2 cells induced survival ability of the cells in suspension culture indicating generation of anoikis resistance. In 3-D cultures KRAS or BRAF mutated Caco-2 cells formed multicellular clusters, partially or completely filled with cells surviving without contact with the extracellular matrix. These observations indicate, that KRAS and BRAF mutations contribute to anoikis resistance and that formation of multicellular clusters in 3-D culture may be a marker of anoikis resistance.

Applying histopathology, we analysed colorectal carcinoma specimens and identified three types of multicellular clusters containing cells without contact with the extracellular matrix, including micropapillary structures, cribriform structures and solid structures. Absence of contact with the extracellular matrix was confirmed with immunohistochemical stainings for basement membrane proteins and survival of the cells was concluded by comparing the counts of apoptotic cells (M30 staining) in cells with and without extracellular matrix contact. Based on these features, such multicellular clusters were identified as putative anoikis resistant populations. Their amount was determined with a digital image analysis. Abundance (high total areal density) of the putative anoikis resistant subpopulations was an independent marker of short cancer specific survival, and areal densities of cribriform and solid structures as such also had similar prognostic effect. Amount of micropapillary structures did not associate to prognosis, yet this subpopulation showed features of quiescence as based on low apoptosis and proliferation rates. This makes them important candidates to explore further since anoikis resistance and tumor dormancy are potentially a combination predicting poor treatment response.

The findings from the studies propose that formation of multicellular clusters containing cells without matrix contact serves as a marker of anoikis resistance both in 3-D cell culture and in actual colorectal carcinomas. Accordingly, it may be possible to evaluate occurrence and even severity of anoikis resistance by conventional histopathological analysis. However, this concept needs confirmatory experiments to show correlation between classical anoikis resistance tests to those based on histopathological features, along with assessment of absence of other ECM proteins from the putative anoikis resistant subpopulations. Finally, convenient recognition of anoikis resistant subpopulations allows better dissection of both genetic and epigenetic background factors and for the analyses of molecular mechanisms of anoikis resistance. In addition, it may be possible to evaluate whether quantitation of anoikis resistance could serve for better treatment selection.