To break or not to break – context matters

ABSTRACT

Oncogene expression can lead to replication stress and genome instability. Recently, we identified oncogene-induced fragile sites (FSs) and revealed that the landscape of recurrent fragility in the same cell type is dynamic. This implies an additional level of complexity in the molecular basis of recurrent fragility in cancer.

Recurrent genome instability is attributed to positive selection and/or sensitivity of specific genomic regions to breakage. Among these are fragile sites (FSs), genomic regions sensitive to replication stress conditions that appear as recurrent breaks on metaphase chromosomes. To date, most mapped FSs have been induced by the DNA polymerase inhibitor aphidicolin and mapped in lymphocytes. In the last decade it has become clear that numerous aphidicolin-induced FSs are cell-type specific: FSs that have been mapped in one cell type are not fragile in another cell type, and FSs shared between cell types can vary significantly in the extent of fragility, being very fragile in one cell type and less fragile in others.¹ An association between cancer breakpoints and FSs was observed a long time ago, and since then has been addressed in many studies.^2-4

More recently it was shown that oncogene expression can lead to replication stress and DNA damage.^5-7 Since FSs are sensitive to replication stress, this discovery can explain the preferential breakage of aphidicolin-induced FSs in human precancerous and cancer lesions.^6,7 However, the cause of most recurrent genome instabilities in cancer remains unknown, since no association with recessive cancer genes or known aphidicolin-induced FSs has been reported.³ Therefore, we were intrigued to investigate whether oncogene expression induces FSs and, if so, to define the repertoire of oncogene-induced FSs and compare it to that of aphidicolin-induced FSs.⁸

Cyclin E and Harvey rat sarcoma viral oncogene homolog (HRAS) are both known cellular oncogenes. Overexpression of cyclin E or activating mutations in HRAS such as G12V are found in many human cancers. Overexpression of both oncogenes induces replication stress, resulting in DNA double-strand breaks and chromosomal rearrangements.^5,6 As expected, we revealed that overexpression of cyclin E or mutated HRAS leads to chromosomal fragility visible on metaphase chromosomes in the form of FSs.

Mapping the cytogenetic locations of FSs induced by aphidicolin, cyclin E, or HRAS in normal diploid BJ-hTERT cells revealed that the repertoire of FSs in the same cell type is dynamic, as the repertoire of FSs induced by oncogenes only partially overlapped with the repertoire of aphidicolin-induced FSs. Moreover, each oncogene induced a unique repertoire of FSs.

The oncogene-induced FSs share prominent features with aphidicolin-induced FSs. Similar to aphidicolin-induced FSs, many oncogene-induced FSs colocalize with recurrent deletion clusters identified in cancer cells, including recurrent deletion clusters that were previously categorized as unexplained. Another feature shared by aphidicolin- and oncogene-induced FSs is colocalization with large genes (>600 kb).

Interestingly, in another recent study Teixeira et al. examined loss of specific genomic regions upon cyclin E deregulation in epithelial cells.⁹ While the 2 regions identified in this paper overlap with cyclin E-induced FSs in BJ-hTERT cells, most regions do not. This supports the notion that oncogene-induced fragility depends on the combination of the specific aberrantly expressed oncogene and the cell type in which it is expressed. In the course of cellular transformation and tumor evolution cells continue to acquire genomic changes, including aberrant oncogene expression and loss of tumor suppressors, due to genomic instability and cellular selection. It is therefore likely that plasticity in the repertoire of FSs also takes place in later stages of cancer development. This may reflect intratumor heterogeneity, which in turn has an impact on clinical outcomes that are often challenged by the evolution of drug resistant clones of cancer cells.

Oncogene overexpression affects various cellular processes including cell metabolism, cell cycle regulation, DNA replication, and transcription. Changes in these processes are likely to underlie the changes in the landscape of fragility (Fig. 1). Hence, we studied the transcriptional status of several extremely large genes (>600 kb) located in genomic regions that we identified as FSs in either cells with cyclin E overexpression or aphidicolin-treated cells, but not in both. Transcriptional changes that correspond to the difference in fragility were found in 2 large genes, Dab, reelin signal transducer, homolog 1 (DAB1) and neurexin 3 (NRXN3). As most identified FSs harbor at least one large gene, it is possible that utilizing more sensitive tools for the detection of transcriptional activity will reveal an even stronger association between large transcription units and fragility.¹⁰ Along with transcriptional changes it is reasonable to assume that other, yet undetermined, processes affected by oncogene overexpression are involved in the variable sensitivity of specific genomic regions to breakage. Especially, changes affecting origin regulation are expected to have implications for fragility and are thus good candidates for further research.

Figure 1.

The landscape of recurrent fragility in a certain cell type is not constant. The repertoires of fragile sites (FSs) induced by chemicals (aphidicolin, an inhibitor of DNA polymerases) or oncogene overexpression (cyclin E or HRAS) only partially overlap. Changes in transcription, the replication program, or yet unknown factors that occur during cellular transformation are likely to be involved in the molecular mechanisms underlying the different sensitivity of genomic regions to replication-induced fragility.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.