A map of human cancer signaling

We conducted a comprehensive analysis of a manually curated human signaling network containing 1634 nodes and 5089 signaling regulatory relations by integrating cancer‐associated genetically and epigenetically altered genes. We find that cancer mutated genes are enriched in positive signaling regulatory loops, whereas the cancer‐associated methylated genes are enriched in negative signaling regulatory loops. We further characterized an overall picture of the cancer‐signaling architectural and functional organization. From the network, we extracted an oncogene‐signaling map, which contains 326 nodes, 892 links and the interconnections of mutated and methylated genes. The map can be decomposed into 12 topological regions or oncogene‐signaling blocks, including a few ‘oncogene‐signaling‐dependent blocks’ in which frequently used oncogene‐signaling events are enriched. One such block, in which the genes are highly mutated and methylated, appears in most tumors and thus plays a central role in cancer signaling. Functional collaborations between two oncogene‐signaling‐dependent blocks occur in most tumors, although breast and lung tumors exhibit more complex collaborative patterns between multiple blocks than other cancer types. Benchmarking two data sets derived from systematic screening of mutations in tumors further reinforced our findings that, although the mutations are tremendously diverse and complex at the gene level, clear patterns of oncogene‐signaling collaborations emerge recurrently at the network level. Finally, the mutated genes in the network could be used to discover novel cancer‐associated genes and biomarkers. Cancer is largely a genetic disease that is caused by acquiring genomic alterations in cells. It is proposed that a malignant tumor arises from a single cell, which undergoes a series of evolutionary processes of genetic or epigenetic changes and selections so that a cell within the population can acquire additional selective advantages, resulting in progressive clonal expansion (Nowell, 1976). Enormous efforts have been made over the past few decades to identify gene mutations that are causally implicated in human cancer. Recently, a whole‐genome or large‐scale efforts toward the identification of genetic and epigenetic changes in tumor samples have been conducted (Stephens et al, 2005; Sjoblom et al, 2006; Greenman et al, 2007; Ohm et al, 2007; Schlesinger et al, 2007; Thomas et al, 2007; Widschwendter et al, 2007). These studies showed that a substantial fraction of the cancer‐associated mutated and methylated genes is involved in cell signaling. Although a wide variety of genetic and epigenetic alterations contribute to the signaling of tumorigenesis, it has been challenging to gain a global view of where and how they affect the signaling alterations to generate tumors on the entire signaling network. To address this question, we performed an integrative analysis of a human signaling network incorporating the cancer mutated and methylated genes. We uncovered an overall picture of the network architecture to determine the sites at which oncogenic stimuli occur and the oncogenic regulatory mechanisms underlying the mutated and methylated genes. Genetic mutations preferentially occur in the proteins (signaling hubs) that receive and send more signals but not in the proteins (neutral hubs) that simply have more physical interactions with others. However, methylated genes have no such preference. Furthermore, we showed that genetic mutations are enriched in positive regulatory loops, whereas methylated genes are enriched in negative regulatory loops. These results suggest that genetic and methylated alterations have different regulatory mechanisms in tumorigenesis. Signaling information propagates through a series of built‐in regulatory motifs to contribute to cellular phenotypic functions (Ma'ayan et al, 2005). The transition from a normal cellular state into an oncogenic state is often driven by prolonged activation of downstream proteins, which are regulated by upstream proteins or regulatory motifs. In cancer cells, constitutive activation of the oncogene signaling is necessary. The enrichment of genetic mutations in positive regulatory loops suggests that the mutants in the motifs must have gain of function or increase their biochemical activities compared with the wild‐type genes to constitutively activate the downstream proteins. Indeed, a recent survey showed that 14 out of the 15 PI3K mutants in tumors have gain of function (Gymnopoulos et al, 2007). A gain‐of‐function mutant in a positive regulatory loop offers the amplification of weak input stimuli and serves as information storage to extend the duration of activation of the affected downstream proteins. This might allow the downstream signaling cascades to persistently hold and transfer information leading to tumor phenotypes. One the other hand, methylation is a known mechanism of inducing loss of function of genes (Ohm et al, 2007; Widschwendter et al, 2007). Negative regulatory loops suppress positive signals and play an important role in maintaining homeostasis and restraining the cellular‐state transitions (Ma'ayan et al, 2005). A loss of function by gene methylation in a negative regulatory loop could inhibit the negative‐feedback mechanism, thereby releasing the restrained activation signals and promoting the oncogenic state transition. Both the gain‐of‐function mutated genes in positive regulatory loops and the loss‐of‐function methylated genes in negative regulatory loops could break this delicate balance, thus promoting state transitions and tumor phenotypes. Extensive efforts have been made to illustrate cancer signaling during the past few decades. However, it has been a struggle to get clues of how the oncogene signaling is structurally and functionally organized. To answer these questions, we extracted an oncogene‐signaling map from the network, which contains 326 nodes, 892 links and the interconnections of mutated and methylated genes (Figure 3). We further systematically identified the ‘oncogene‐signaling‐dependent events’ (the phenomenon by which certain cancer cells become dependent on certain signaling cascades for growth or survival), which are frequently used in many tumors. Within the map, the oncogene‐signaling‐dependent events form three highly connected regions that resemble oncogene‐signaling superhighways that are frequently used in tumorigenesis (Figure 3). Two of the regions consist of genes that are heavily methylated in cancer stem cells. This map provides a blueprint of the oncogene signaling in cancer cells and can be used to generate testable hypotheses for a given mutation in a particular cancer sample. To get insights into how the map is functionally organized, we first divided the map into 12 oncogene‐signaling blocks based on the connectivity of the map nodes. We then queried the 592 tumor samples, in which each sample contains at least two mutations of the network genes, using the 12 oncogene‐signaling blocks. Interestingly, we found that two oncogene‐signaling blocks are enriched in gene mutations and tend to collaborate in most tumor types (Figure 4A). These two blocks are called p53 (composed of p53, p14, Rb, BRAC1 and BRAC2 etc.) and Ras (Ras, PI3K and EGFR etc.) blocks. In all the tumor types analyzed, at least 2 signaling gene mutations, one from the p53 block and the other from another block, are necessary for tumorigenesis and further support the notion that both the prevention of cell death (p53 block) and the promotion of cell proliferation (Ras or other blocks) are necessary to generate most tumors. The same analysis was extended to six representative cancer types. Breast and lung cancers have more complex oncogene‐signaling block collaborative patterns than other four cancer types that have similar oncogene‐signaling block collaborative patterns found in the 592 samples. We further benchmarked the gene mutation data from the systematic sequencing of tumor samples using the oncogene‐signaling map as a framework. We also obtained a oncogene‐signaling block collaborative pattern similar to that found in the 592 tumor samples, when using the mutation data of the NCI‐60 cancer cell line, in which 24 known cancer genes were screened for mutations (Figure 4B). For the data derived from the genome‐wide mutation screening, colon cancer showed a simple collaborative pattern of the oncogene‐signaling blocks, whereas breast tumors showed complex patterns (Figure 4C and D). These findings imply that, although the mutations seem tremendously diverse and complex at the gene level, clear patterns emerge recurrently at the network level in most tumors. This work uncovered novel features of human cancer signaling that help in understanding the underlying mechanisms of tumorigenesis. Furthermore, it provides a conceptual and technical framework for incorporating tumor genome sequencing outputs to get more insights into the cancer‐signaling mechanisms that will lead in identifying the key genes for biomarkers and drug development.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.