Friday 29 August 2008
The spectrum of genomic losses ranges from cytogenetically visible alterations, such as complete or partial chromosomal monosomies (large-scale genomic losses), to small regional deletions or single-gene or intragenic deletions that are detectable only by techniques that provide high spatial resolution (small genomic losses).
Most recurrent genomic losses probably contribute to malignant transformation by reducing the function of specific genes in the affected chromosomal regions.
Since restoration of gene function is more challenging than, for example, inhibition of increased kinase activity, it is unclear whether direct pharmacologic targeting of genomic losses will ever be possible. Nevertheless, an improved understanding of the functional consequences of these aberrations may lead to the identification of indirect targets for therapeutic intervention.
For example, inactivation of the PTEN tumor-suppressor gene on band 10q23.3, which occurs with high frequency in glioblastoma, prostate cancer, and endometrial cancer, increases signaling through the phosphoinositide-3-kinase–AKT–mammalian target of rapamycin (PI3K–AKT–mTOR) pathway and promotes tumor-cell proliferation and survival. Experimental models and early clinical trials indicate that PTEN-deficient tumors are sensitized to the growth-suppressive activity of mTOR inhibitors, such as sirolimus (also called rapamycin).
The dissection of the mechanisms through which genomic losses promote tumorigenesis is challenging. Recent developments include the application of modern genomic techniques to the study of large-scale genomic losses, the identification of new tumor-suppressor genes that act through allelic insufficiency, and the discovery of noncoding genes as functionally relevant targets of recurrent genomic losses.
Large-Scale Genomic Losses
Extensive genomic deletions affecting multiple genes are frequent in tumors, making it difficult to identify which lost gene contributes to the development of the cancer.
The classic approach to identifying a tumor-suppressor gene compares multiple tumors with a specific chromosomal deletion to determine the minimal genomic region that is lost in all cases. Candidate genes from this region are then screened for deletions, mutations, or epigenetic modifications that inactivate the remaining allele.
For many recurrent genomic losses, however, such as 1p deletions (Ip LOH) in neuroblastoma, 3p deletions (3p LOH) in lung cancer, and 7q deletions (7q LOH) in myeloid cancers, the critical genes are unknown.
Regardless of whether the respective disease genes have been identified, some deletions have proved to be of great value for determining the prognosis and guiding treatment decisions, as exemplified by 5q deletions (5q LOH) in acute myeloid leukemia; 11q deletions (11q LOH), 13q deletions (13q LOH), and 17p deletions (17p LOH) in chronic lymphocytic leukemia; and the concurrent 1p deletions (1p LOH) and 19q deletions (19q LOH) in anaplastic oligodendroglioma.
New genomic techniques have considerably improved the identification of functionally relevant genes within regions of recurrent chromosomal deletions.
For example, RNA interference screening in combination with high-resolution DNA copy-number analysis identified the REST gene as a suppressor of epithelial-cell transformation that maps to a segment of band 4q12 that is frequently deleted in colon cancer.
The power of array-based SNP genotyping as a tool for gene discovery in cancers associated with genomic losses is demonstrated by recent studies that revealed deletions of PAX5 (9p13) and IKZF1 (7p13-p11.1) in approximately 30% of children with B-progenitor acute lymphoblastic leukemia and in more than 80% of patients with BCR-ABL1–positive acute lymphoblastic leukemia, respectively.
Genomic Losses Resulting in Allelic Insufficiency
Another difficulty in the analysis of chromosomal deletions occurs in the identification of genes that contribute to tumorigenesis by inactivation of a single allele.
Since such haplo-insufficient tumor-suppressor genes cannot be identified through analysis of the remaining allele, alternative approaches are required to assess the consequences of monoallelic deletion.
An example is a recent study in which graded down-regulation of multiple candidate genes by RNA interference was used to identify RPS14 as a causal gene for the 5q minus syndrome, a subtype of the myelodysplastic syndrome characterized by a 1.5-Mb commonly deleted region on chromosome band 5q32.
Notably, patients with the 5q minus syndrome are highly responsive to the thalidomide derivative lenalidomide,88 although the mechanisms through which lenalidomide restores normal erythropoiesis remain unknown.
Monoallelic deletions can completely inactivate tumor-suppressor genes that are located on the X chromosome because humans carry only one functional copy of all X-linked genes.
This mechanism was documented in a recent study that identified small deletions of band Xq11.1, targeting the FAM123B tumor-suppressor gene, in 21.6% of patients with sporadic Wilms’ tumors.
DNA sequence analysis subsequently identified additional patients with inactivating FAM123B mutations, again highlighting the potential of chromosomal imbalances for guiding the discovery of alternative genetic changes with similar functional consequences.
Genomic Losses Affecting Noncoding Genes
Cancer-associated chromosomal losses may act through inactivation of genes that do not encode proteins. For example, several genomic regions that are recurrently deleted in a variety of tumors contain microRNA genes.
These genes encode small RNAs involved in post-transcriptional regulation of gene expression, and there is growing evidence that the loss of specific microRNAs with tumor-suppressive activity may contribute to tumorigenesis.
This pathogenetic mechanism was shown by the observation that MIRN15A and MIRN16-1 are located within a segment of band 13q14.3 that is deleted in approximately 50% of patients with chronic lymphocytic leukemia and the subsequent discovery that MIRN15A and MIRN16-1 negatively regulate the expression of the antiapoptotic protein BCL2.
Given that many chromosomal regions that are recurrently deleted in cancer appear to lack protein-coding genes that normally act to limit cell proliferation, it seems plausible that the analysis of cancer-associated genomic losses will reveal additional tumor-suppressor microRNAs.