Emerging Potential of Therapeutic Targeting of Ubiquitin-Specific Proteases in the Treatment of Cancer Free

Most of the current molecular cancer therapies target protein kinases and mediate their antitumor activity through the deactivation of an aberrantly expressed, tumor-dependent, or unregulated target enzyme. Clinically successful kinase-targeting examples include Herceptin (Her2), imatinib (Bcr-Abl and c-Kit), vemurafenib (BRAF), and ibrutinib (BTK; ref. 1). However, kinase inhibition has restricted application and limited long-standing efficacy, as intrinsic resistance reduces benefit to only a fraction of patients and most responders will acquire resistance due to additional mutations or activation of compensating pathways (1). A deeper understanding of the molecular pathways associated with cancer has implicated posttranslation regulation through ubiquitination/deubiquitination of many target proteins as an additional and attractive targeted therapeutic approach.

Covalent attachment of ubiquitin, a 76–amino acid protein, to a target protein is a means of regulating protein half-life, localization, and activity. Because protein homeostasis is essential for the survival of all cells, but more essential to cancer cells, modulation of individual ubiquitin–proteasome system (UPS) components may present an opportunity for therapeutic targeting. As proof-of-principle, many compounds with proteasome inhibitory activity have been developed, including bortezomib (or Velcade; Millenium Pharmaceuticals), which is a synthetic dipeptide boronic acid that reversibly inhibits the chymotryptic-like activity of the 20S enzymatic core of the proteasome and induces apoptosis in several malignancies. Velcade is approved by the U.S. Food and Drug Administration (FDA) for the treatment of patients with mantle cell lymphoma and multiple myeloma, even those resistant to doxorubicin, mitoxantrone, melphalan, and dexamethasone, and is commonly used in combination with many of these agents (2). Amplified protein synthesis (immunoglobulin) in many myeloma cells may underlie their clinical sensitivity to bortezomib and other proteasome inhibitors as most solid tumors do not have a similar commitment to elevated protein synthesis and are not clinically responsive to these drugs. This narrow therapeutic application, combined with some toxicity (sensory neuropathy), may be circumvented by novel proteasome inhibitory molecules (2).

Targeted inhibition of ubiquitin-conjugating enzymes and ligases could provide another therapeutic avenue. Inhibition of NEDD8-associated E1 enzyme by MLN4924, E2 enzyme hCdc34 by CC0651, and E3 ligase MDM2 by RITA (NSC652287) and MI-219 reflects this ongoing effort (3). Deubiquitinases (DUB) are another class of emerging anticancer target that regulate specific substrate proteins by reversing their ubiquitination through the hydrolysis of isopeptide or α-peptide bonds linking ubiquitin to the target protein (4). If the target protein is an oncogene, the DUB activity of the associated DUB may stabilize its cellular expression and supports identification of DUB inhibitors that could reestablish normal protein turnover, location, or activity (4). This approach may also avoid the deleterious side effects associated with direct targeting of the proteasome. With around 55 members, ubiquitin-specific proteases (USP) comprise the largest and most diverse family of DUBs. Genetic and/or functional analysis have placed USPs in the category of cancer-associated proteases, and their unique biochemical structures have made them desirable targets for anticancer therapies. More than 30 USPs have been associated with cancer directly or indirectly. These numbers are not surprising, considering the various critical cellular functions regulated by different USPs and the diversity of substrates used and regulated by them. A comprehensive list of USPs altered in different cancers is provided in Table 1.

Recurrent mutations of USPs are rare in cancer with the exception of CYLD. Germline mutations of the tumor-suppressor gene CYLD are prevalent in familial cylindromatosis, a genetic condition that leads to predisposition for developing multiple skin tumors (5). The only known chromosomal translocation involving a USP is the fusion of the promoter of CDH11 to the full-length USP6 gene that leads to upregulated transcript levels of USP6 (6). USP gene fusions are reported for USP42 and USP16 that are fusion gene partners with RUNX, which has high-frequency gene alterations in hematologic diseases such as chronic myelomonocytic leukemia and acute myeloid leukemia (AML; refs. 7, 8). USP18 transcript levels are upregulated in mice expressing the AML fusion protein AML1–ETO (9). USP33/VDU1 mRNA levels are overexpressed in B-cell acute lymphoblastic leukemias (10). Increased mRNA levels of USP50 have also been reported in AML (11). Upregulated transcript levels of USP9x have been correlated with poor prognosis in multiple myeloma and are associated with Mcl-1 accumulation in patients with follicular lymphoma (12). Thus, USPs are repeatedly implicated in many hematologic malignancies.

An important aspect of hematopoietic tumors is that p53 mutations are rare; however, p53 levels are tightly regulated by ubiquitination/deubiquitination (13). Thus, modulation of p53 levels in these tumors can provide a therapeutic advantage by inducing p53-dependent apoptosis and cell-cycle arrest. A key molecule involved in p53 ubiquitination is the E3 ligase MDM2, which is a negative regulator of p53 stability (13). Although USP7 and USP2a have been reported to deubiquitinate MDM2, leading to degradation of p53 and antiapoptotic activity, USP7 and USP10 can directly deubiquitinate p53 under specific conditions (14, 15). Under normal conditions, USP7 stabilizes MDM2 and leads to degradation of p53, whereas USP10 deubiquitinates and stabilizes p53. In response to DNA damage and following ATM kinase activation, the substrate preference of USP7 switches from MDM2 to p53, leading to its stabilization (14, 16). Similarly, in response to DNA damage, USP10 is stabilized in an ATM kinase-dependent manner and a portion of it translocates to the nucleus where it joins USP7 in deubiquitinating and stabilizing p53. Other USPs known to regulate p53 levels in the cell include USP4 and USP5 (17, 18). USP4 reduces p53 levels in the cell through direct binding and deubiquitination of the E3 ligase HUWE1, also called MULE or ARF-BP1 (17). USP5 depletion in the cell leads to accumulation of unanchored polyubiquitin chains, competing with p53 for proteasomal destruction, which results in nuclear accumulation of p53 and transcriptional activation of p53 target genes (18). Very recently, USP15 was shown to target MDM2 with effects on the stability of both p53 and the T-cell transcription factor, NFATc. Thus, USP15 inhibition could result in direct tumor cell apoptosis and increased T-cell responsiveness (19).

Because ubiquitination/deubiquitination is also an important mechanism essential for regulation of the DNA-damage repair (DDR), a large number of USPs are implicated in regulating or coordinating DDR. USP3, USP7, USP10, USP11, USP16, USP21, and USP22 are reported to regulate DDR sensor proteins. USP1, USP2, USP4, USP7, USP10, USP11, USP24, USP29, and USP47 are directly implicated in the regulation of DDR repair proteins (20). Single-strand DNA breaks are repaired by nucleotide excision repair (NER) and base excision repair (BER) pathways that deal with various DNA helix-distorting lesions and single-strand breaks; mismatch repair pathways that repair base mismatches and insertions/deletions; the nonhomologous end-joining pathway (NHEJ) and/or homologous recombination (HR) pathways, and the Fanconi anemia (FA) pathway, which in conjunction with certain HR factors act to recognize and repair inter-strand cross-links (ICL) DNA lesions. USP1 and USP7 act in multiple DDR pathways. USP1 removes monoubiquitin from FANCD2 and PCNA, and thus regulate ICL and trans-lesional DNA synthesis (20). USP1 levels are well regulated during the cell cycle and their depletion leads to genomic instability. USP1 knockouts show defects in FA and HR repair (21). Like USP1, USP7 also acts in multiple DDR pathways. In response to DSB, the DSB sensor kinase ATM is activated, which leads to downregulation of USP7S, a USP7 isoform, resulting in the activation of a p53 response and regulation of the G₂–M checkpoint (20). USP7 is also a major regulator of the oxidative stress response. Through its deubiquitinating activity, USP7 can modulate FOXO4 transcriptional activity in response to oxidative stress. USP7 can also modulate BER of oxidative lesions through chromatin remodeling directly by deubiquitination of histones, or indirectly by regulating the cellular levels of E3 ubiquitin ligases involved in histone ubiquitination. USP29 also mediates the oxidative stress response by stabilizing p53 in a FBP/JTV1–dependent manner. USP7 is also implicated in transcription-coupled NER (TCNER), a subpathway of NER that efficiently removes highly toxic RNA polymerase II blocking lesions on DNA causing Cockayne syndrome and UV-sensitive syndrome (UVSS). Apart from USP7, USP47 is the only other DUB that is reported to regulate BER through the stabilization of newly synthesized polymerase β (Pol-β). USP47-silenced cells show aberrant DDR growth inhibition and chemosensitization. USP3 and USP16 deubiquitinate histone H2A and USP16 also opposes the RNF8–RNF18 pathway–mediated DDR to double-strand breaks (DSB), leading to accumulation of replication stress in the cell. USP24 stabilizes DDB2, also called p48, which is an important protein in the recognition and repair of UV-induced lesions in the NER pathway (22). In response to DNA damage, USP28 binds with 53BP1 and its loss induces ionizing radiation–induced apoptosis. USP44 has been identified as a DUB critical for spindle assembly checkpoint activity and in centrosome regulation. USP44-deficient mice are susceptible to spontaneous tumors of the lung, supporting a role for USP44 as a tumor-suppressor gene (reviewed in ref. 20). These findings are especially noteworthy because current radiation and chemotherapy regimens used for the treatment of various cancers rely heavily on inducing DNA damage to the cancer cell. Developing specific or even partially specific inhibitors that can target several USPs, resulting in cell-cycle arrest or apoptosis, in the cancer cell may be advantageous.

USPs are also implicated as important regulators of several cancer-associated pathways. The NF-κB pathway is constitutively activated in a vast range of cancers and is implicated in tumorigenesis and metastasis. Several USPs are reported as negative regulators of NF-κB signaling. CYLD binds to the NEMO component of the IκB kinase complex and regulates its activity through deubiquitination of TRAF2. Inhibition of CYLD increases resistance to apoptosis, an effect that can be relieved by inhibiting NF-κB activity using aspirin derivatives (23). USP31 has also been identified as a TRAF2-interacting protein that regulates NF-κB activation by members of the TNF receptor superfamily (24). USP4 directly interacts and exerts deubiquitinating activity on multiple NF-κB pathway–associated molecules, such as TRAF2, TRAF6, and TAK1 kinase, and negatively regulates TNFα and IL1β-induced cancer cell migration. USP4 negatively regulates receptor-interacting protein 1 (RIP1)–mediated NF-κB activation and promotes TNFα-induced apoptosis in head and neck squamous cell carcinomas through direct interaction with RIP1 and deubiquitination of K63-linked ubiquitin from RIP1 (25). USP11 mediates downregulation of TNFα-mediated NF-κB activation through modulation of IκBα stability by its deubiquitination (26). USP14 removes the ubiquitin chain of I-κB, thereby inducing I-κB degradation and increasing cytokine release in lung epithelial cells (27). COP9 signalosome (CSN) regulates assembly and activity of cullin-RING ubiquitin ligases (CRL), which are involved in the ubiquitination of Iκ-α. USP15 is identified as a CSN-associated DUB that reverses the ubiquitinating activity of CSN on IκBα, thus negatively regulating NF-κB signaling (28).

Several USPs are implicated in the regulation of TGFβ signaling at various levels of the pathway. USP4, USP11, and USP15 interact with the TGFβ receptor I, deubiquitinate, and stabilize this receptor to engage sustained Smad activation, resulting in enhancement of TGFβ signaling. Inhibiting USP4 and USP11 inhibits TGFβ-mediated epithelial–mesenchymal transition (EMT) and invasion in breast cancer while USP15 inhibition blocks EMT and invasion in glioblastomas (29). USP4 has also been shown to drive cross-talk between TGFβ and AKT signaling in breast cancer. Inhibiting USP4 suppresses AKT-mediated breast cancer cell migration. These findings underscore the relevance of USP4 in the pathway as AKT activation has been associated with poor prognosis in breast cancer (29). USP9X reverses the mono-ubiquitination of Smad4 at Lys 519, a modification that prevents its binding to phosphorylated Smad2 and inhibition of TGFβ signaling. Thus, Smad4 deubiquitination by USP9X reinstates TGFβ signaling (12). USP9X can also deubiquitinate E3 ubiquitin ligase SMURF1, whose substrates include TGFβ receptor and coreceptor Smads (30). Recent findings in our laboratory show an oncogenic potential of USP9X in established breast cancer cells and inhibiting USP9X inhibits cell growth in tumorigenic cells and induces apoptosis in triple-negative breast cancers (Unpublished Data). Thus, there is accumulating evidence in favor of therapeutic targeting of USPs like USP4, USP15, and USP9X in breast cancer.

Inhibitor of differentiation (ID) proteins (ID1-4) regulate differentiation and maintain stem cell fate through inhibition of basic–helix–loop–helix transcription factors. ID protein levels are deregulated during cancer and have been associated with an aggressive clinical phenotype and poor patient outcomes. The expression levels of ID proteins are tightly regulated through ubiquitination by the APC/Cdh1 E3 ligase complex, which leads to proteasomal degradation and shorter half-life of ID proteins in the cell (31). Recently, USP1 was shown to stabilize ID1, ID2, and ID3 protein expression levels in normal and malignant stem cell populations through its deubiquitinating activity. USP1 was found to be overexpressed in a subset of osteosarcomas where it stabilized ID1, ID2, and ID3, leading to repression of p21 levels and aberrant osteogenic differentiation (32). USP1 abundance correlated with ID2 protein abundance levels in human osteosarcoma tumors. Thus, small molecule–mediated targeting of USP1 may provide an additional modality for treatment of cancer through modulation of ID proteins. Despite the growing interest in USPs and in acquiring USP inhibitors, their development has been hampered by some of the biochemical properties of the USPs themselves and mostly unavailable three-dimensional (3D) structures. To date, only seven USPs have defined 3D structures (Table 1). USPs can range in size from 350 to 3,500 amino acids. They comprise a central catalytic domain, which could be 295 to 850 amino acids long and shows an average 22% identity among USPs and many different additional domains like the ubiquitin-like domain in USP7 and USP14. Such versatility in this small family of DUBs could be a boon in the development of partially selective or specific inhibitors and will benefit from unraveling the structural information of more USPs.

Among the earliest reported DUB inhibitors are cyclopentenone prostaglandins that induce apoptosis and also increase the cellular content of poly-ubiquitinated proteins, suggesting they are nonselective DUB inhibitors (3). A small molecule discovered in screens for DUB inhibitors, PR-619, is selective for DUBs over other cysteine proteases, but inhibits all DUBs tested with moderate potency. More recently, HBX 41,108, which was originally reported to be a USP7 inhibitor, was confirmed to be a nonselective DUB inhibitor. HBX 41,108 stabilizes p53 in HEK293 cells and induces caspase-3 and PARP cleavage in both p53^+/+ and p53^−/− HCT-116 cells. The specificity of HBX 41,108 is limited as it can also inhibit USP5, USP8, UCH-L3, and caspase-3 with a potency (70–200 nmol/L) greater than its activity against USP7 (530 nmol/L). However, because of the critical regulatory functions played by USP7 in the cell there is tremendous interest in developing USP7-specific inhibitors (33). Progenra has identified the novel P5091 small-molecule USP7-specific inhibitor that can stabilize p53, inhibit cancer cell proliferation, and is an active antitumor agent in various tumor models. P5091 induces apoptosis in multiple myeloma cells resistant to conventional and bortezomib therapies (34). P5091 is well-tolerated in animals, inhibits tumor growth, prolongs survival, and triggers synergistic anti–multiple myeloma activity in combination with other chemotherapeutic agents such as lenalidomide, HDAC inhibitor, or dexamethasone. Hit-to-lead optimization identified additional analogs of P5091 (e.g., aqueous soluble derivative P045204) that increases the steady-state levels of p53 and its transcriptional target p21 in a time-dependent manner in HCT-116 cells.

b-AP15, also known as VLX1500, is a unique class of proteasome inhibitor that inhibits the activity of the 19S regulatory particle–associated DUBs, UCHL5, a ubiquitin C terminal hydrolase, and USP14. b-AP15 induces apoptosis in tumor cells irrespective of their mutant p53 status and BCL2 overexpression and is effective in inhibiting tumor progression in multiple solid tumor mouse models and dissemination of an in vivo AML model. The apoptotic effects of b-AP15 are mediated through the induction of oxidative and endoplasmic reticulum (ER) stress in response to b-AP15–mediated accumulation of poly-ubiquitinated proteins (35).

Pimozide and GW7647 inhibit USP1–UAF1 complex through noncompetitive binding, which provides for more selectivity and specificity toward USP1–UAF1. These inhibitors act in synergy with cisplatin to inhibit cell proliferation in cisplatin-resistant non–small cell lung cancer cells (36). Pimozide also inhibits leukemic cell growth through degradation of USP1 substrate ID1 (37). ML323 is another highly potent inhibitor of the USP1–UAF1 complex that provides for excellent selectivity against DUBs. In non–small cell lung cancer and osteosarcomas, ML323 potentiates cisplatin-induced cytotoxicity (38). ML323-mediated inhibition of USP1 inhibits deubiquitination of FANCD2 and PCNA and compromises TLS and FA pathways (38). Thus, ML323 may provide a means to sensitize cancer cells to platinum-based therapies.

In our laboratory, we described and developed the small-molecule inhibitor WP1130, also known as Degrasyn, which was derived from a compound with JAK2 kinase–inhibitory activity. WP1130 rapidly induces accumulation of poly-ubiquitinated proteins, resulting in induction of apoptosis. WP1130 is a partially selective inhibitor that directly inhibits the deubiquitinating activity of USP9X, USP5, and USP14, all of which regulate survival protein stability and proteasome function. In several studies done in the recent past, USP9X inhibition by WP1130 promotes apoptosis by reducing MCL-1 levels and increased tumor cell sensitivity to chemotherapy (39). Recently, USP9X inhibition by WP1130 was shown to inhibit the growth of ERG-positive tumors in vitro and in mouse xenograft models of prostate cancer (40). Our laboratory has been actively engaged in identifying derivatives with greater selectivity, activity, and drug-like properties. We recently identified and are developing a compound with improved specificity toward USP9X and antitumor activity in mouse models of myeloma, lymphoma, and melanoma (unpublished data).

Conclusively, there is accumulating experimental evidence that a large number of USPs could be targeted for anticancer therapeutics. Early evidence for antitumor efficacy with the currently available USP inhibitors is more than encouraging and sets the stage for the development of selective, as well as partially selective, small-molecule DUB inhibitors.

No potential conflicts of interest were disclosed.