Previously we found that the cytoplasmic domain of MUC1 (MUC1-CD) induces multiple, stable transcriptional changes in transfected cells. We investigated these changes in in vivo models and identified two unique MUC1-CD-dependent expressional signatures. One, which we denoted as the MTS (MUC1 Tumorigenesis Signature), was identified in vivo and reflected the interactions of tumor cells with the host microenvironment, as was evidenced by the activation of genes involved in angiogenesis and extracellular matrix signaling . A second expressional signature representing lipid and cholesterol metabolism was also identified in the context of tumor-stromal interactions . Therefore, in the current report, we focused on the detection of MUC1-CD-dependent transcriptional changes unique to oncogenic cells. We hypothesized that these "intrinsic" changes would be important for MUC1-CD signaling leading to oncogenesis and may be connected with the fundamental mechanisms of the malignant transformation. In this regard, we profiled MUC1-CD-transformed cells grown in vitro, without any influences from the host stroma.
In the current report, we describe a MUC1-Associated Proliferation Signature (MAPS) that provides independent prognostic information, adding to standard pathological evaluation and clinical staging of lung adenocarcinoma. This signature was derived from the set of genes initially detected in an experimental system as upregulated by MUC1-CD in vitro and potentially involved in a highly oncogenic phenotype . The MAPS is distinct from our previously reported MUC1-related signature (MTS, ) and was identified using a contrasting experimental approach. The major functional groups of differentially expressed genes in vitro represented cellular growth, proliferation and cell cycle control. In comparison, the MTS was derived from genes highly enriched by functional groups representing cell motility, metastasis and angiogenesis. We believe that these results demonstrate important differences between the intrinsic properties of tumor cells and the properties that are determined by tumor-stromal interactions. Interestingly, only two genes (CDC20 and RRM2) were common between both signatures, perhaps indicating that, at least in our experimental system, expression of these two genes is independent of the host microenvironment.
All the genes that comprise the MAPS are related to cell cycle control and proliferation. For instance, CDC20 (homolog of S. cerevisiae cell division cycle 20 protein) directly binds to and activates anaphase-promoting complex (APC), which leads to ubiquitination and degradation of cyclin B (CCNB1) and therefore promotes the onset of anaphase and mitotic exit . The APC/CDC20 complex is under negative control of MAD2L1 (human homolog of S. Cerevisiae MAD2) and BUB1 (see Table 1). Also, PRC1 (protein regulating cytokinesis 1) is a human homolog of S. cerevisiae Ase1, which is involved in spindle formation and also promotes anaphase and mitotic exit . CDC2 (cell division cycle 2), or CDK1 (cyclin-dependent kinase 1), is a catalytic subunit of a protein kinase complex, called the M-phase promoting factor, formed with cyclin B1 (CCNB1) that induces entry into mitosis . CDC2 phosphorylates securin, which is another target of APC/CDC20 and is an inhibitor of separase-protease, responsible for the cleavage of sister chromatid cohesions. CDC2-dependent phosphorylation of securin protects it from APC/CDC20-induced ubiquitination and degradation . CDKN3 (cyclin-dependent kinase inhibitor 3) is a dual-specificity protein phosphatase that interacts with CDC2 and CDK2 and inhibits their activity . These data show that six of the seven genes comprising MAPS not only belong to a cell cycle-related functional group but represent a specific pathway of interacting proteins responsible for anaphase control, chromosome segregation and mitotic entrance/exit (see also Figure 1B). RRM2 (ribonucleotide reductase, M2 subunit) encodes the small subunit (R2) of ribonucleotide reductase, the heterodimeric enzyme that catalyzes the rate-limiting step in deoxyribonucleotide synthesis. Using siRNA screening, Kittler et al.  identified 37 genes required for cell division, one of which was RRM2.
There is substantial literature indicating that the genes in MAPS are co-expressed and are involved in tumorigenesis and cancer progression. Five of seven MAPS genes are upregulated in immortalized breast cancer cell lines compared to primary breast tumor cell cultures (CDC2, CDC20, CDKN3, MAD2L1 and RRM2)  and all seven MAPS genes are upregulated in response to infection of HPV-18, a virus associated with cervical cancer, in keratinocytes . All seven were also found to be co-expressed with E2F, which is expressed in breast cancer compared with normal breast tissue and is elevated during the G2/M transition . This suggests a possible role for E2F inhibitors in treating poor-prognosis cancers that express MAPS. All seven MAPS genes are downregulated in response to Brd4 transfection in a mouse mammary cell line and are included in a 141-gene prognostic signature based on differential expression in this cell line. Expression of this signature correlated with prognosis in five separate human breast cancer cohorts . This is one of many published results from tumor expression profiling experiments which have linked increased expression of genes from common pathways involved in cell growth and proliferation to poor outcomes in cancer patients . A meta-signature was identified consisting of sixty-nine genes expressed more in high-grade compared to low-grade tumors in eight separate microarray analyses spanning seven types of cancer including lung adenocarcinoma . These included many genes associated with cell proliferation, including five of the seven genes in our abbreviated MAPS signature: CCNB1, CDC2, CDC20, CDKN3 and MAD2L1. Thus, MAPS reflects a pattern of gene expression associated with high-grade cancers, but having greater prognostic significance than histological grade in our results.
Current data of Whitfield et al.  indicate that proliferation-associated genes can be considered not only as common prognostic/predictive markers in different cancers, but also as promising targets for anti-cancer therapy. Among the genes comprising MAPS, at least two are targets of known drugs. These are hydroxyurea for RRM2 and flavopiridol and staurosporin (UCN-01) for CDK1 (CDC2). In addition, taxol, which affects microtubule formation and blocks mitosis at the G2/M transition, may have interactions with 6 of 7 gene products included in MAPS. RRM2 is also a target that may be used to potentiate chemotherapy. Kittler et al.  demonstrated that silencing of RRM1 and RRM2, which encode the large and small subunits of the human ribonucleotide reductase (RNR) complex, respectively, markedly enhanced the cytotoxicity of the topoisomerase I (Top1) inhibitor camptothecin (CPT). Silencing of RRM2 was also found to enhance DNA damage as measured by γ-H2AX. Upregulation in RRM2 expression levels suggests an active role for RNR in the cellular response to DNA damage that could potentially be exploited as strategy for enhancing the efficacy of Top1 inhibitors . The MUC1-CD is also involved in the control of DNA damage response . The data presented in the current report suggest that this control may be associated with a set of genes regulating G2/M transition and exit from mitosis through the network of reactions connected with spindle formation and chromosome segregation.
Many existing biomarkers that have been identified for non-small-cell lung cancer indicate the presence of disease, as in screening or recurrence. The genes in the MAPS are, however, potential biomarkers of prognosis and could help guide treatment in patients with a new diagnosis of primary lung cancer. There are cytokeratin biomarkers that have been studied which show evidence of prognostic significance in lung adenocarcinoma, including CYFRA 21-1, TPA and TPS . These biomarkers are detected on the protein levels in relatively high concentrations from freshly prepared tissues. Our signature has the potential to be measured by PCR at picogram levels both in frozen tissues and paraffin-embedded archival samples. Further prospective investigations are needed to compare potential protein and RNA-based biomarkers, which might be complementary to each other.