|Home | About | Journals | Submit | Contact Us | Français|
E3 ubiquitin ligases are a large family of proteins that are engaged in the regulation of the turnover and activity of many target proteins. Together with ubiquitin-activating enzyme E1 and ubiquitin-conjugating enzyme E2, E3 ubiquitin ligases catalyze the ubiquitination of a variety of biologically significant protein substrates for targeted degradation through the 26S proteasome, as well as for nonproteolytic regulation of their functions or subcellular localizations. E3 ubiquitin ligases, therefore, play an essential role in the regulation of many biologic processes. Increasing amounts of evidence strongly suggest that the abnormal regulation of some E3 ligases is involved in cancer development. Furthermore, some E3 ubiquitin ligases are frequently overexpressed in human cancers, which correlates well with increased chemoresistance and poor clinic prognosis. In this review, E3 ubiquitin ligases (such as murine double minute 2, inhibitor of apoptosis protein, and Skp1-Cullin-F-box protein) will be evaluated as potential cancer drug targets and prognostic biomarkers. Extensive study in this field would lead to a better understanding of the molecular mechanism by which E3 ligases regulate cellular processes and of how their deregulations contribute to carcinogenesis. This would eventually lead to the development of a novel class of anticancer drugs targeting specific E3 ubiquitin ligases, as well as the development of sensitive biomarkers for cancer treatment, diagnosis, and prognosis.
The ubiquitin-proteasome pathway is a major pathway for the targeted degradation of proteins and involves multistep enzymatic reactions catalyzed by a cascade of enzymes, including ubiquitin-activating enzyme E1, ubiquitin-conjugating enzyme E2, and ubiquitin ligase E3. Ubiquitin is first activated by binding to E1 through a thioester bond between a cysteine residue at the active site of E1 and the C-terminus glycine (G76) of ubiquitin. Activated ubiquitin in an E1-ubiquitin complex is then transferred to E2, which also forms a thioester bond between its active-site cysteine residue and the G76 of ubiquitin. Finally, ubiquitin is covalently attached to the target protein through an isopeptide bond between the G76 of ubiquitin and the ε amino group of an internal lysine residue of the target protein, catalyzed by E3 ubiquitin ligase. Through multiple runs of reactions, ubiquitin is covalently attached to substrates to form K48-linked polyubiquitinated conjugates that are rapidly recognized and degraded by the 26S proteasome . Recent data have shown that proteins can also be monoubiquitinated or polyubiquitinated through K63 linkage, leading to altered protein activity and subcellular localization, rather than degradation [2,3]. A diagram of ubiquitination reaction and the three potential fates of proteins after ubiquitination are illustrated in Figure 1.
E3 ubiquitin ligase is an enzyme that binds to specific protein substrates and promotes the transfer of ubiquitin from a thiolester intermediate to amide linkages with proteins or polyubiquitin chains . Because they serve as the specific substrate-recognition element of the system, E3 ligases play an important role in ubiquitin-mediated proteolytic cascade. There are approximately 1000 E3 ligases in the human genome that can be classified into three major types, based on their domain structure and substrate recognition. The first class comprises N-end rule ubiquitin ligases that target protein substrates bearing specific destabilizing N-terminal residues, including Arg, Lys, His (type I), and Phe, Trp, Leu, Tyr, and Ile (type II) . One recent example of protein degradation by the Ub-dependent N-end rule pathway is Drosophila inhibitor of apoptosis protein (IAP) . The second type of E3 is HECT, with the first family member being E6-associated protein (E6-AP), which, together with oncoprotein E6, promotes p53 ubiquitination and degradation . HECT E3 ligases contain an approximately 350-amino acid C-terminal region homologous to that of E6-AP, with a conserved active-site cysteine residue near the C-terminus, through which HECT domain E3 ligases form thioester intermediates with Ub [8,9]. N-terminal regions are highly variable and may be involved in substrate recognition . The third and largest type of E3 ligase is the Really Interesting New Gene (RING) family, which contains a classic C3H2C3 or C3HC4 RING finger domain  with a characteristic linear sequence of Cys-X2-Cys-X9–39-Cys-X1–3-His-X2–3-Cys/His-X2-Cys-X4–48-Cys-X2-Cys, where X can be any amino acid. A RING finger domain binds to two zinc atoms per molecule in a cross-braced system, where the first and third pairs of cysteine/histidine form the first binding site and where the second and fourth pairs of cysteine/histidine form the other .
E3 ubiquitin ligases exist and act as a single peptide [such as murine double minute 2 (Mdm2) and X-linked IAP (XIAP)] or as multiple component complexes [such as Skp1-Cullin-F-box protein (SCF)]. Through the covalent modification of a vast repertoire of cellular proteins with ubiquitin, E3 ubiquitin ligases regulate almost all aspects of eukaryotic cellular functions and biologic processes. Accumulating data have strongly suggested that deregulation of E3 ligases contributes to cancer development and that overexpression of E3 ligases is often associated with poor prognosis [12,13]. Thus, E3 ligases, which determine the specificity of protein substrates and are themselves “drugable” enzymes, can serve as potential cancer targets as well as cancer biomarkers.
An ideal cancer target meets the following criteria: 1) it plays an essential role in cancer genesis, and/or is required for the maintenance of cancer cell phenotype, and/or is apoptosis-protective and confers cancer cells resistance to apoptosis; 2) it is overexpressed in cancer cells, and its overexpression is associated with a poor prognosis of patient survival; 3) inhibition of its expression or activity induces growth suppression and/or apoptosis in cancer cells; 4) it is “drugable,” meaning that it is an enzyme (e.g., kinase) or a cell surface molecule (e.g., membrane-bound receptor) that can be easily screened for small-molecule inhibitors or that can be targeted by a specific antibody; and 5) most importantly, it is not expressed or is expressed at a very low level in normal cells, and its inhibition has a minimal effect on normal cell growth and function. Thus, inhibition of such a target would achieve a maximal therapeutic index with minimal toxicity. The E3 ubiquitin ligases discussed below would meet some of these criteria.
Mdm2 encodes a 90-kDa protein that was first identified as the gene responsible for the spontaneous transformation of an immortalized murine cell line BALB/c 3T3 . It contains a p53-binding domain at the N-terminus and a RING domain at the C-terminus. The p53-binding domain of Mdm2 or Hdm2 (human counterpart of Mdm2) binds to the tumor suppressor p53, whereas the RING domain acts as an E3 ubiquitin ligase to promote rapid degradation of p53 [15–19]. Both in vitro and in vivo studies indicated that oncogen ic activity of Mdm2 is mainly attributable to its binding and degradation of p53 [20,21]. p53 is a classic tumor suppressor that is inactivated in more than 50% of human cancers. Under unstressed conditions, the p53 level is very low in cells due to Mdm2 binding and degradation. On DNA damage by ionizing radiation or anticancer drugs, p53-Mdm2 binding is dissociated as a result of p53 phosphorylation and acetylation, leading to p53 activation. Activated p53 acts as a transcription factor that transactivates a number of genes, leading to growth arrest (to repair damaged DNA) or apoptosis (if damage is too severe to repair) .
Because Hdm2 plays a critical role in the regulation of p53 level, Hdm2 appears to be a promising cancer target whose inhibition should lead to p53 reactivation and apoptosis induction in human cancer cells harboring wild-type p53. Indeed, Hdm2 as a cancer drug target has been extensively validated. The approaches used include: 1) the blockage of Hdm2-p53 interaction with synthetic peptides or monoclonal antibodies [23,24]; and 2) the reduction of Hdm2 levels with antisense oligonucleotides or siRNA [25–27]. As consequence of Hdm2 targeting, the levels of p53 increased, followed by transactivation of p53 downstream target genes and induction of growth arrest and apoptosis [23–27]. All these proof-of-concept studies support the notion that Hdm2 appears as a cancer target for the discovery and development of a new class of small molecular antagonists. Indeed, two classes of small molecules that target Mdm2 to reactivate p53 and to induce apoptosis have been discovered so far: the first class comprises inhibitors of Mdm2 E3 ubiquitin ligase and the second class comprises compounds that disrupt Mdm2-p53 binding, as illustrated in Figure 2.
A family of small molecules—HLI98 series—was identified through an HTS of a chemical library of 10,000 compounds using an in vitro Hdm2 autoubiquitination assay . Follow-up experiment showed that the compound indeed inhibited Hdm2 activity, as well as other E3 ligases and even E2 ligases at higher concentrations. In cell-based assays, the compound stabilized p53 and Mdm2, and activated p53-dependent transcription and apoptosis, but also had p53-independent cytotoxicity. Furthermore, as expected, the compound worked much better in cancer cells containing wild-type p53 than in those containing mutant p53 because targeting Hdm2 should, in theory, have little or no effect on human cancers with mutant p53. However, in vivo antitumor activity of HLI98, using human xenograft models, has not been reported . Nevertheless, this proof-of-concept study indicated that Hdm2 E3 is a valid cancer target and that it is possible to identify more potent inhibitors of Hdm2 E3 ligase as a novel class of anticancer drugs for future discovery and development.
Three classes of structurally distinctive compounds, namely, Nutlin, RITA (Reactivation of p53 and Induction of Tumor cell Apoptosis), and a nonpeptide Mdm2 inhibitor (MI-17), were reported to disrupt Hdm2-p53 binding [29–31]. Historically, it has been difficult to develop small-molecule inhibitors to disrupt large protein-protein interactions. However, the crystal structure of Mdm2-p53 peptide binding revealed that binding relies on the contact of the p53 peptide side chains of Phe19, Trp23, and Leu26 with the N-terminus of Mdm2 (amino acids 17–125) in a deep hydrophonic pocket , which made it possible for small molecules to disrupt binding. Indeed, the Nutlin series was identified through a screening of a diverse library that disrupted Mdm2-p53 peptide binding , whereas structure-based design on Mdm2-p53 binding pocket led to identification of a potent nonpeptide Mdm2 inhibitor MI-17 with a chemical structure different from that of Nutlin . Conversely, RITA was identified through a cell proliferation assay using a pair of isogenic cancer cell lines differing in p53 status. RITA bound to p53 and prevented p53-Hdm2 interaction . Compared to Hdm2 E3 ligase inhibitors, Mdm2-p53 binding inhibitors appeared to be much more potent and specific in activating the p53 pathway, leading to growth arrest, apoptosis, and in vivo tumor growth inhibition. Again, mechanistically, these compounds will only work in human cancers harboring wild-type p53 and, preferentially, with Mdm2 overexpression. This certainly turned out to be the case [29–31].
Two major concerns are associated with these compounds, which induce p53 accumulation through Mdm2 manipulation. The first concern is the therapeutic window or the selectivity between normal and cancer cells. Although it is still unclear mechanistically, the activation of p53 in normal cells by these compounds mainly caused growth arrest, rather than apoptosis, making it possible to achieve a therapeutic window by adjusting the dose regime and the duration of the treatment  (S. Wang, personal communication). The second concern is the oncogenic activity of Mdm2 independent of p53. Because Mdm2 itself is a p53 target, p53 activation, as a result of either approach, would cause significant accumulation of Mdm2. An increased amount of even ligase-deficient Hdm2 might actually promote tumor growth . This potential side effect needs to be further addressed. Finally, in addition to Mdm2, two additional RING proteins, COP1  and PIRH2 , were determined to be p53 targets and to promote p53 ubiquitination and degradation. Indeed, COP1 and PIRH2 were overexpressed in a subset of human cancers with increased p53 ubiquitination [36,37]. Further validation of COP1 and PIRH2 as promising cancer targets is a prerequisite to initiating a library screen for their specific inhibitors.
The IAP family has at least eight members, including XIAP, cIAP-1, cIAP-2, Ts-IAP, NAIP, survivin, Livin/ML-IAP, and Apollon/Bruce [38–41]. They all contain one or several baculovirual IAP repeat (BIR) domains that are required for the suppression of apoptosis. Some family members also have a RING finger domain at the C-terminus for the ubiquitination and degradation of caspases [39,42,43]. In XIAP, BIR3 (the third BIR domain) potently inhibits the activity of the active caspase-9, whereas the linker region between BIR1 and BIR2, as well as the BIR2 domain itself, selectively targets active caspase-3 or caspase-7 [44,45]. Thus, IAP suppresses apoptosis by binding to and by inhibiting active caspase-3, caspase-7, and caspase-9 through BIR domains [38,39,46]. In apoptotic cells, caspase inhibition by IAP is negatively regulated by a mitochondrial protein, second mitochondria-derived activator of caspase (Smac). Smac physically interacts with multiple IAPs and relieves their inhibitory effect on caspase-3, caspase-7, and caspase-9. Smac binds to the BIR3 domain of XIAP through four N-terminal residues (AVPI) that recognize a surface groove on BIR3. These four amino acids are conserved in three Drosophila proteins (Reaper, Grim, and Hid) that induce apoptosis by eliminating the binding of Drosophila IAP to caspases [47,48].
IAP as a promising cancer target has been extensively validated by overexpression, silencing, or the use of a Smac-derived AVPI peptide that binds to IAPs to free up caspases. Indeed, overexpression of IAP suppressed apoptosis induced by a variety of stimuli [39,49], whereas downregulation of XIAP or survivin through antisense RNA or siRNA has been shown to induce apoptosis in many human cancer cell lines [50–57]. Furthermore, Smac peptide consisting of AVPI sequence sensitized many human cancer cells to apoptosis induced by conventional cancer therapies both in vitro and in vivo [58–61], indicating that it is feasible to identify AVPI-like small molecules to disrupt IAP-caspase binding. The current effort, therefore, was focused on IAP-caspase binding inhibitors. Analogous to the Mdm2-p53 case, the use of small molecules to disrupt protein-protein binding was made possible by a well-defined small binding packet between the IAP BIR3 domain and the AVPI peptide of Smac [47,48]. At least five classes of such compounds have been discovered so far, and their structures have been summarized in a recent review article .
The first class comprises tripeptides of unnatural amino acids that were developed through a structure-based design targeting the BIR3 domain of XIAP. The compounds induced apoptosis in a number of human cancer cell lines by releasing active caspase-9 from XIAP binding [63,64]. The structure-based computational screening of a three-dimensional structure database of traditional herbal medicines led to the discovery of embelin as a potent inhibitor of XIAP-caspase-9 binding . Embelin activated caspase-9, inhibited cell growth, and induced apoptosis in prostate cancer cells with high levels of XIAP, with minimal effect on normal prostate epithelial and fibroblast cells, containing low levels of XIAP expression .
The next two classes comprise compounds targeting BIR2 or BIR2 link regions to disrupt binding to caspase-3. Aryl sulfonamide, identified through the biochemical screening of a combinatorial chemical library, disrupted XIAP-caspase-3 interaction and sensitized cancer cells to the activator of the death receptor pathway . A polyphenylureas series was identified with chemical library screening using an enzyme derepression assay by overcoming XIAP-mediated suppression of caspase-3 . These series of compounds indeed increased caspase activity, had broad activity against cancer cell growth as tested in 60 NCI cancer cell lines, sensitized cancer cells to chemotherapeutic drugs, and inhibited tumor cell growth in xenograft models in mice with limited toxicity to normal tissues . This class of compounds has been shown recently to activate caspase-3 and caspase-7 and to directly induce the apoptosis of leukemia cell lines and primary samples from acute myelogenous leukemia patients without much lethal effect on normal hematopoietic cells .
Finally, a Smac mimic small-molecule compound was identified through structure-based design using computer-simulated conformations of AVPF as a guide. The compound bound to XIAP, c-IAP1, and c-IAP2 to activate caspase-3 and sensitized cancer cells to apoptosis induced by TNF[tumor necrosis factor]-related apoptosis-inducing ligand (TRAIL) and tumor necrosis factor α . We further evaluate the compound in three breast cell lines with various levels of IAP. Acting alone, Smac mimic compound was quite potent with a cytotoxic IC50 of 3.8 nM in high IAPs expressing MDA-MB-231 cells, but was inactive at a much higher concentration in low IAPs expressing T47D and MDA-MB-453 cells. In fact, as low as 2.5 nM Smac mimic alone was sufficient to activate caspase-3 and to induce apoptosis in MDA-MB-231 cells. In combination treatments with TRAIL or etoposide, Smac mimic significantly sensitized cells to growth suppression and apoptosis in MDA-MB-231 cells, but to much lesser extent in T47D and MDA-MB-453 cells. Thus, in these cell lines, Smac mimic compound acts in an apparent IAP-dependent manner to induce apoptosis alone, as well as to sensitize breast cancer cells to TRAIL- or etoposide-induced apoptosis through caspase-3 activation .
All these studies showed convincingly that IAPs are valid cancer targets and that disrupting their binding to caspases would release active caspases to induce apoptosis preferentially in IAPs overexpressing human cancer cells, with less toxicity on normal cells having low IAP expression. Conversely, IAPs have been shown to act as ubiquitin ligases to promote the ubiquitination and degradation of caspase-3, caspase-9, and Smac [71–74]; mutations of the RING domain, which are required for E3 ligase activity, reduced the apoptotic activity of XIAP . Thus, targeting their ubiquitin ligases appears to be a feasible approach to increasing the levels of caspases and Smac, thus inducing apoptosis in cancer cells or sensitizing cancer cells to conventional cancer therapies. It is of concern that even ligase activity is inhibited, however, IAPs may still bind to caspases and prevent caspase activation and apoptosis induction. Thus, specific inhibitors of IAP E3 ligase, which are yet to be discovered, would be more effective in combination therapy with chemotherapeutic drugs or IAP-caspase binding inhibitors. Figure 3 illustrates IAP targeting to activate caspases and to induce apoptosis.
SCF and SCF-like complexes comprise the largest family of E3 ubiquitin ligases that consist of Skp1, Cullins, F-box protein, and ROC/Rbx/SAG (Sensitive to Apoptosis Gene) RING finger protein [75,76]. The crystal structure of the SCF-Rbx complex revealed that Cul-1 acts as a scaffold that binds Skp1-F-boxSkp2 (the protein substrate-recognition complex; at its N-terminus) and Rbx1 (which recruits E2; at its C-terminus) . Thus, SCF E3 ubiquitin ligases may serve as scaffolds that position substrates and E2 enzyme optimally for ubiquitin transfer. Through various F-box proteins acting as substrate receptors , SCF ligases recognize many protein substrates and promote their ubiquitination and degradation, thus regulating a variety of biologic processes [75,76]. For example, through Skp2, which binds to cyclin-dependent kinase inhibitors p27, p21, and p57, SCF E3 promotes their ubiquitination and degradation, thus promoting G1→S progression [79–81]. Skp2 also binds to c-Myc to promote its ubiquitination and degradation and, at the same time, acts as a coactivator to enhance c-Myc-induced S-phase transition and to activate c-Myc target genes [82,83]. Through the F-box protein β-TrCP, SCF E3 ligase promotes the ubiquitination and degradation of Emi-1 (early mitotic inhibitor), an inhibitor of the anaphase-promoting complex, to control meiotic and mitotic progression [84,85]. β-TrCP also binds to IκB and β-catenin and, together with other components of SCF ligase, promotes their ubiquitination and degradation [86–94], thus regulating NFκB and Wnt signaling pathways. Accumulated evidence strongly suggested that abnormal regulations of SCF E3 ubiquitin ligase contribute to uncontrolled proliferation, genomic instability, and cancer .
The validation of whether SCF E3 ubiquitin ligase is an appealing cancer target has been mainly focused on its components, particularly Skp2, β-TrCP, and SAG, using either overexpression or silencing (through antisense or siRNA) approaches. Skp2 overexpression in gastric carcinoma cells decreased the level of p27, increased cell growth rate, rendered cancer cells more resistant to actinomycin D-induced apoptosis, and increased their invasion potential . Tissue-specific expression of Skp2 in the prostate gland of a mouse transgenic model caused significant downregulation of p27 level and marked tissue overproliferation, leading to hyperplasia, dysplasia, and low-grade carcinoma . Skp2, when targeted for expression in T-lymphoid lineage, cooperated with activated N-Ras to induce T-cell lymphomas with a short latent period and high penetrance, indicating that Skp2, as a protooncogene, is involved in the pathogenesis of lymphomas . Conversely, downregulation of Skp2 using an antisense oligonucleotide remarkably suppresses the growth of small cell lung cancer cells . siRNA silencing of Skp2 has been shown to inhibit the growth of melanoma cells , oral cancer cells , glioblastoma cells , and lung cancer cells [102,103]. Similarly, overexpression of β-TrCP increased NFκB activity and chemoresistance, whereas silencing of β-TrCP by siRNA reduced NFκB activation and chemoresistance in pancreatic cancer cells . Transgenic mice with β-TrCP1-targeted expression in the intestine, liver, and kidney had an increased incidence of tumor formation in these organs . Conversely, silencing β-TrCP1 through siRNA or overexpression of dominant-negative mutant was shown to suppress the growth and survival of human breast cancer cells . In the case of SAG, the second member of ROC/Rbx family and a RING component of SCF [76,107], its overexpression protected cells and tissues from apoptosis induced by redox reagents and by ischemia/reperfusion-generated hypoxia in a RING domain-dependent manner [107–110]. SAG overexpression also promotes cell growth under serum-starved conditions , whereas antisense SAG transfection inhibits tumor cell growth .
Although substantial progress has been made in this area, SCF E3 ligase as a cancer target suffers from several intrinsic drawbacks. The first is specificity. The same SCF E3 ligase can promote the degradation of either oncogenes or tumor-suppressor genes, dependent on different F-box proteins or even the same F-box protein . The therapeutic outcome of ligase inhibitors has to be cell context-dependent, which is hard to manage in cancer patients without a thorough understanding of the mechanism. Secondly, assay complexity is a big issue, although several screening assays for E3 ligase inhibitors have been developed recently . SCF is a multiple component E3 ligase, and its intrinsic enzymatic mechanism is still unclear, except that the core ligase components Rbx1/ROC1-Cullins have been shown to promote autoubiquitination in an in vitro assay [114–116]. Indeed, it is feasible to identify general inhibitors against Cul-ROC/SAG core E3 ligase using an in vitro assay described for the inhibitor screening of APC2/APC11 core ligase ; such inhibitors would not, however, have a desired specificity against any particular SCF complex. It is uncertain to conduct high-throughput screening (HTS) using all SCF E3 components because, unlike kinases or proteases, SCF ligase does not contain an evident central enzymatic active site to which small molecules could bind. A three-dimensional structure-based computer design strategy has been proposed to assess whether interfaces among SCF components are suitable for small-molecule binding . An alternative approach is to screen for inhibitors that disrupt binding between SCF components. One example is the development of an HTS assay for inhibitors of Cks1-Skp2 interaction that would lead to p27 accumulation [119,120]. Figure 4 illustrates two potential approaches to target SCF-Skp2 E3 ligase with the expected outcome of p27 accumulation and growth inhibition. Nevertheless, it appears that SCF E3 ligase itself may not be a practical target per se. However, its components may serve as cancer biomarkers for further development and use in cancer clinics (see below).
Early diagnosis and treatment of cancer would significantly improve the survival of cancer patients. The development of cancer biomarkers for early detection or prognosis prediction is of significant importance. An ideal cancer biomarker will meet some criteria for a cancer target (e.g., high expression in cancer tissues, but not in normal tissues, with a causal relationship with cancer genesis, development, or metastasis). In addition, it should be a secretory protein that is readily obtained and identified from a patient's body fluids, such as serum, urea, stool, and sputum. For intracellular biomarkers to which E3 ubiquitin ligases belong, it should be overexpressed or it should have a high frequent mutation rate so that it can be readily identified by immunohistochemical or mutational analyses using tumor tissues from biopsy. Traditional biomarkers of cell proliferation, such as Ki-67 and PCNA, have had a mixed clinical track record . New and more reliable cancer markers are being searched and developed . Due to the overexpression of some E3 ubiquitin ligases in a number of human cancers with associated poor prognosis [13,123], it is possible that these E3 ligases (such as Hdm2 and the F-box protein Skp2) can be further characterized and developed as useful cancer biomarkers.
Although it is normally expressed at a low level, Hdm2 is overexpressed through gene amplification, increased transcription, or enhanced translation in a variety of human cancers, including breast carcinomas, soft tissue sarcomas, esophageal carcinomas, lung carcinomas, glioblastomas, and malignant melanomas [123–125]. Among the 28 human tumor types examined, the overall frequency of Hdm2 gene amplification is 7%, with the highest frequency being observed in soft tissue tumors (20%) . Furthermore, high Hdm2 levels are often associated with poor prognosis, with an increased likelihood of distant metastases and with a poor response to therapeutic drugs [124,126]. Due to the important role of Hdm2 in promoting the degradation of tumor suppressor p53, the development of Hdm2 as a cancer biomarker, particularly in soft tissue sarcomas, is highly desirable.
The expression and prognostic significance of RING-containing IAPs, such as XIAP, c-IAP1, and c-IAP2, have been extensively studied in human tumor and cells lines with mixed results. The study with 60 NCI cancer cell lines revealed that higher levels of XIAP or c-IAP1 proteins correlated with sensitivity or resistance to some chemotherapeutic drugs, respectively. Conversely, acute myeloid leukemia (AML) patients with lower levels of XIAP proteins had a significantly longer survival, with a tendency toward a remission longer than that of patients with higher levels of XIAP . Similarly, high expression levels of XIAP correlated with poor overall survival in childhood de novo AML . In another study, however, expression levels of XIAP had no prognostic impact on AML patients . In radically resected non-small cell lung cancer patients, a high XIAP predicted a longer overall survival . In cervical carcinoma, the basal expression levels of IAPs had no prognostic significance . In clear cell renal carcinomas, a significant inverse correlation was achieved between XIAP expression and tumor aggressiveness, and patients' survival . Furthermore, survivin, a RING-less IAP, was found to be overexpressed in most common human cancers, but not in normal terminally differentiated adult tissues. The resistance of cancer cells to conventional cancer therapy and a worse clinical prognosis are usually correlated with a high expression of survivin [39,49,133–136]. Due to these mixed results, it is unlikely that IAPs, with the probable exception of survivin, will be developed as useful cancer biomarkers.
It has been well documented that Skp2 acts as an oncogene mainly by targeting p27 for ubiquitination and degradation . Overwhelming evidence showed that Skp2 is overexpressed in almost all major human cancers, including carcinomas of the breast, colon, lung, brain, prostate, and liver, among many other human cancers. In most cases, Skp2 overexpression is inversely correlated with p27 expression and is directly correlated with poor clinic prognosis (for a review, see Nakayama and Nakayama ). Thus, Skp2-p27 inverse correlation may deserve further characterization for clinical use as a prognostic index.
Among other SCF components, expression of β-TrCP1 was found to be elevated in colon cancers (particularly in those with metastases) , pancreatic carcinomas , and hepatoblastomas . Overexpression of β-TrCP2 was also detected in primary prostate, breast, and gastric cancers, as well as in the cell lines derived from these cancers [140–142]. Cul4A was recently found to be involved in Mdm2-mediated p53 degradation , as well as in DDB1-Skp2-mediated p27 degradation . Cul4A gene was amplified and overexpressed in primary breast cancers  and hepatocellular carcinomas . Finally, SAG was overexpressed in a subset of colorectal carcinomas  and in non-small cell lung cancers . Importantly, high SAG expression was correlated with poor patient survival and could serve as a useful prognostic marker .
E3 ubiquitin ligases regulate a variety of biologic processes, including cell growth and apoptosis, through the timely ubiquitination and degradation of many cell cycle- and apoptosis-regulatory proteins. Abnormal regulation of E3 ligases has been convincingly shown to contribute to cancer development [12,13,148]. Thus, targeting E3 ubiquitin ligases for cancer therapy has gained increasing attention, which is further stimulated by the recent approval of a general proteasomal inhibitor, bortezomib (Velcade, Millennium, MA), for the treatment of relapsed and refractory multiple myeloma , as well as for the discovery of a new class of proteasome inhibitors . In contrast to general proteasome inhibitors, targeting a specific E3 would selectively stabilize a specific cellular protein regulated by a particular E3, thus avoiding some unwanted effects on other cellular proteins. This would, therefore, achieve a high level of specificity with less (at least in theory) associated toxicity. Because several HTS assays are now in place for the rapid screening of small molecular inhibitors of E3 ligases , it is anticipated that, in the near future, specific inhibitors of E3 ubiquitin ligase will be discovered and developed as a novel class of anticancer drugs. E3-based cancer biomarkers will be also developed and used in clinics as diagnostic tools or prognostic indices for the benefit of cancer patients.
1This work is supported by National Institutes of Health grant R01-CA111554.
2I apologize to those whose works were not cited due to space limitations.