Targeting HAUSP in both p53 wildtype and p53-mutant tumors

ABSTRACT

Inhibition of Mdm2 function is a validated approach to restore p53 activity for cancer therapy; nevertheless, inhibitors of Mdm2 such as Nutlin-3 have certain limitations, suggesting that additional targets in this pathway need to be further elucidated. Our finding that the Herpesvirus-Associated Ubiquitin-Specific Protease (HAUSP, also called USP7) interacts with the p53/Mdm2 protein complex, was one of the first examples that deubiquitinases (DUBs) exhibit a specific role in regulating protein stability. Here, we show that inhibitors of HAUSP and Nutlin-3 can synergistically activate p53 function and induce p53-dependent apoptosis in human cancer cells. Notably, HAUSP can also target the N-Myc oncoprotein in a p53-independent manner. Moreover, newly synthesized HAUSP inhibitors are more potent than the commercially available inhibitors to suppress N-Myc activities in p53 mutant cells for growth suppression. Taken together, our study demonstrates the utility of HAUSP inhibitors to target cancers in both a p53-depdentent and -independent manner.

Introduction

Tumor suppressor p53 can be activated by myriad stimuli to control a broad network of barriers to prevent cancer growth—apoptosis, senescence, and ferroptosis, to name a few [1]. Therefore, it is not surprising that p53 is highly controlled at all levels; the major system to regulate p53 protein stability is through balancing ubiquitination and deubiquitination [2]. Many E3-ubiquitin ligases have been identified for p53, but the most prominent is the oncoprotein Mdm2 [3]. Mdm2 amplification, which drives p53 degradation and promotes tumorigenesis, occurs in about 30% of human sarcomas that retain wildtype p53 alleles [4]. Conversely, knockout of murine Mdm2 drastically stabilizes p53 protein levels causing p53-mediated embryonic lethality [3]. As a therapeutic strategy to activate p53, independently of broadly stimulating general genotoxic stress, small molecule antagonists were identified to block this p53/Mdm2 interaction. Blocking the p53-binding pocket of Mdm2 decreases ubiquitinated-p53 which stabilizes and activates total p53 protein levels and induces tumor inhibition in an Mdm2-amplified setting [5].

Deubiquitinating enzymes reverse the degradation-specific signals added by E3-ubiuitin ligases by removing ubiquitin moieties from their specific substrates. In the context of p53, our early work identified the first bone fide deubiquitinase of p53 as Herpesvirus-Associated Ubiquitin-Specific Protease (HAUSP, also called USP7) [6]. We showed that HAUSP interacts with and stabilizes p53 protein levels through deubiquitination activity. Interestingly, partial knockdown of HAUSP resulted in slight decreases in p53 levels [7]. Conversely, complete knockdown or genetic knockout of HAUSP stabilized p53 protein levels [7,8]. We and others showed that HAUSP indeed also binds and regulates Mdm2 ubiquitination [7] as well as Mdmx (a negative regulator of p53 without any intrinsic E3-ligase activity) ubiquitination [9]. Therefore, ablation of HAUSP will destabilize the major E3-ubiquitin ligase Mdm2 leading to a subsequent decrease in ubiquitinated-p53 and increase in total p53 protein levels (reviewed in detail [10]). These works, along with studies identifying the complex secondary network of antagonists and activators for HAUSP to modulate either p53 or Mdm2 ubiquitination [10], make targeting HAUSP an attractive therapeutic option for cancers maintaining a wildtype p53 status.

To physiologically understand the function of HAUSP, we generated both conventional [11] and conditional [12] Hausp mutant mice. In both studies, our work indicated that inactivation of HAUSP destabilized Mdm2 levels which caused p53 stabilization. Interestingly, unlike the Mdm2 mutant mice whose phenotype was completely rescued by p53 ablation [13,14], the lethality observed from the Hausp mutant mice were not rescued by p53 inactivation [11,12]. This indicates that although HAUSP is a major regulator of the p53 axis, there must be p53-independent substrates of HAUSP that are controlling embryonic development. Indeed, many researchers have identified novel substrates of HAUSP independent of the p53 axis. Recently, we identified a substrate of HAUSP partially explaining the p53-independent embryonic defects observed in the Hausp mutant mice. Our work showed that the proto-oncogene N-Myc, which controls embryonic development and neuroblastoma cancer survival (in an N-Myc amplified setting) [15], is tightly control by the enzymatic activity of HAUSP, namely through deubiquitination [16]. Conversely, HAUSP knockdown destabilized N-Myc protein levels and controls N-Myc function [16]. Importantly, as N-Myc amplification is correlated with severity and overall survival in a subset of neuroblastomas [15], we demonstrated that inhibition of HAUSP decreases neuroblastoma cell survival and tumorigenesis by destabilizes N-Myc protein levels both in vivo and xenograft models [16].

Similar to the rationale of Mdm2 inhibitors, HAUSP inhibitors were recently developed to target p53 wildtype cancers [10,17–22]. Considering the previously defined role of HAUSP in cancer (reviewed in detail [10]), HAUSP inhibition has been shown to reduce tumor burden in tumors that retain wildtype p53 (reviewed in detail [10]). Notably, our work was one of the first to demonstrate that HAUSP inhibitors could additionally function in a p53-independent manner to prevent tumorigenesis by regulating N-Myc protein levels [16]. Although these inhibitors function as a proof-of-concept that the enzymatic activity of HAUSP can be pharmacologically targeted, there is an increasing need to develop more potent, more soluble, and less toxic HAUSP inhibitors [10].

Here, we show that HAUSP inhibitors are able to stabilize and activate p53. We demonstrate that the combination of low doses of HAUSP inhibitors with low doses of the Mdm2 inhibitor Nutlin-3 [5] was able to synergize in stabilizing p53 levels. Single treatment with either low dose agents was unable to induce apoptosis; notably, the combination of HAUSP and Mdm2 inhibition which strikingly activated p53 levels, also induced apoptosis. Lastly, we demonstrate, using newly generated thiazole derivatives of HAUSP inhibitors [23], an increased potency of HAUSP inhibition in a p53-independent manner.

Results

HAUSP inhibition stabilizes and activates p53 protein levels

The dynamic regulation of HAUSP to modulate p53 stability is well-established. We used isogenic colorectal carcinoma HCT116 p53^+/+ and p53^−/− cells stably expressing a doxycycline-inducible HAUSP-specific short hairpin RNA (shRNA) or GFP shRNA [24] to reconfirm this regulation as a principle for testing HAUSP inhibitors. As expected, upon reduction of HAUSP levels through induction of HAUSP-shRNA, p53 protein levels were stabilized and subsequently activated (as indicated by p21 levels); as a negative control we did not observe any p21 induction in HCT116 p53^−/− cells after HAUSP ablation (Figure 1a). Using native HCT116 p53^+/+ and p53^−/− cells, we tested the commercially available HAUSP inhibitor, HBX41108 [17]. As expected, and similar to the knockdown data, p53 was stabilized and activated when the enzymatic function of HAUSP was inhibited (Figure 1b).

Figure 1.

HAUSP and Mdm2 inhibitors combine to strongly regulated p53 stability. a) Extracts of HCT116 p53^+/+ or isogenic HCT116 p53^−/− cells stably transduced with a doxycycline-inducible shGFP or shHAUSP were cultured in the presence of doxycycline for 48h and immunoblotted with antibodies specific for HAUSP, p53, p21 and Actin. b) Extracts of HCT116 p53^+/+ or isogenic HCT116 p53^−/- cells treated with DMSO (-) or 10µm HAUSP inhibitor HBX41108 (+) for 24h and immunoblotted with antibodies specific for HAUSP, p53, p21 and Actin. c) Extracts of HCT116 p53^+/+ or isogenic HCT116 p53^−/- cells treated with DMSO (-), 5µm Nutlin-3, or 10µm HBX41108 (+) as indicated for 8h and immunoblotted with antibodies specific for HAUSP, p53, p21 and Vinculin. d) Extracts of HCT116 p53^+/+ cells treated with DMSO (-),5µm (+) or 10µm (++) HBX41108, or 5µm (+) or 20µm (++) Nutlin-3 as indicated for 12h and immunoblotted with antibodies specific for cleaved Parp1 and Actin. e) Extracts of HCT116 p53^+/+ cells treated with DMSO (-),5µm Nutlin-3 (+), or 5µm HAUSP inhibitor p22077 (+) as indicated for 36h and immunoblotted with antibodies specific for p53 and Vinculin.

HAUSP inhibition combines with Mdm2 inhibition to stabilize and activate p53 protein levels

Both Mdm2 and Mdmx are direct substrates of HAUSP [7,9]. Knockdown or ablating HAUSP function, leads to an increase of ubiquitinated Mdm2 driving its proteasomal degradation. At the point where Mdm2 levels decrease below a certain threshold, the residual Mdm2 available to ubiquitinate p53 is insufficient and p53 protein levels are subsequently stabilized and activated. Therefore, to exploit this phenomenon, we asked the question whether inhibiting the interaction between the residual Mdm2/p53 after HAUSP inhibition could combine to further stabilize p53 protein levels. To this end, we treated native HCT116 p53^+/+ and p53^−/− cells with a combination of HAUSP inhibitors and Nutlin-3, a small molecule inhibitor of the Mdm2/p53 interaction [5]. As shown in Figure 1c, a single treatment with Nutlin-3 was sufficient to stabilize and activate p53 levels to a similar extent as HBX41108 (lane 2 vs 3). Strikingly, the combination of HBX41108 with Nutlin-3 led to a dramatic stabilization of p53 protein levels (Figure 1c; lane 4). While higher treatment doses of the HAUSP inhibitor HBX41108 (10 µm) induces apoptosis, as indicated by cleaved parp 1, low doses of HBX41108 (5 µm) did not induce apoptosis (lane 3 vs 2; Figure 1d). In contrast neither a low dose (5µm) nor a high dose (20 µm) of Nutlin-3 induced apoptosis after 12 hours (lane 4–5; Figure 1d). Strikingly, the combination of low dose HBX41108 and low dose Nutlin-3 was able to induce apoptosis (lane 6; Figure 1d). Further, a similar trend of p53 stabilization was observed using a separate commercially available HAUSP inhibitor p22077 [18] when combined with Nutlin-3 (lane 4; Figure 1e). Taken together, these data suggest that the residual Mdm2 not destabilized by HAUSP inhibition could be targeted to more potently and efficiently activate p53 protein levels and prevent tumor growth.

Novel HAUSP inhibitors better modulate HAUSP function

Due to poor solubility and low potency of commercially available HAUSP inhibitors, we recently described the synthesis of a series of thiazole derivates [23] based on p5091 [20] and p22077 [18] inhibitors. We selected two lead compounds C1 and C5 to test the potency compared with p5091 (Figure 2a). Previously, we showed that HAUSP binds and stabilizes the proto-oncogene N-Myc in neuroblastoma; conversely, ablation or pharmacological inhibition of HAUSP destabilized N-Myc protein levels and inhibited tumor growth [16]. To reconfirm these findings, we tested the N-Myc amplified neuroblastoma cell line SK-N-DZ. In a dose-dependent manner, HAUSP inhibitor p5091 was able to efficiently destabilize N-Myc protein levels (lanes 2–5; Figure 2b) by inhibiting the enzymatic activity of HAUSP [16]. Lastly, to compare our novel HAUSP inhibitors C1 and C5 with p5091, we treated SK-N-DZ cells with two doses of each inhibitor. As shown in Figure 2c, both C1 and C5 was able to more efficiently destabilize N-Myc protein levels at a low dose (lane 2 and 4 vs 6). These data indicate that our novel HAUSP inhibitors are more potent than the commercially available p5091.

Figure 2.

Novel HAUSP inhibitors regulate N-Myc stability. (a) Chemical structure and formula of commercially available HAUSP inhibitor p5091 and thiazole derivative compounds C1 and C5. (b) Extracts of SK-N-DZ cells treated with DMSO or a serial titration of p5091 from 25 µm-3.125 µm for 8 h and immunoblotted with antibodies specific for HAUSP, N-Myc, and Actin. (c) Extracts of SK-N-DZ cells treated with DMSO or increasing amounts (6.25 µm–12.5 µm) of C1, C5, and p5091 for 8 h and immunoblotted with antibodies specific for HAUSP, N-Myc, and Actin.

Discussion

By reversing ubiquitin-mediated degradation signal on substrates, deubiquitinases have the ability to regulate many if not all disease pathways. Although challenging, the generation of small molecule inhibitors targeting different deubiquitinating enzymes has gained much interest in the field and many are even advancing preclinical development [25]. Therefore, dissecting and understanding of context and substrate specificity for each deubiquitinase will greatly accelerate this process. In regards to HAUSP, the vast amount of research from identifying the structure of the catalytic domain, to the multiple identified substrates, and higher levels of HAUSP regulation has eased the testing of newly generated HAUSP inhibitors.

In this work, we reiterate the regulation of HAUSP inhibition on p53 stabilization. We show that HAUSP inhibitor HBX41108 can stabilize and activate p53 protein levels (Figure 1b). Interestingly, combining a low dose of HAUSP inhibitor (either HBX41108 or p22077) with the Mdm2 inhibitor Nutlin-3 strikingly stabilizes p53 protein levels (Figures 1c and 1e). This suggests that the remaining pool of Mdm2 not destabilized by HAUSP inhibition, can be targeted to further stabilize p53. Importantly, this stark activation of p53 was sufficient to induce apoptosis, which is in contrast to low dose single agent treatment of either inhibitor (Figure 1c). These results further suggest that acute and low dose combination treatment may be more effective than increasing the dose of either single agent, which may result in more unwanted off-target toxicity. Lastly, we further characterize our newly synthesized HAUSP inhibitors [23] in a p53-independent context. We show in Figure 2c, that both C1 and C5 HAUSP inhibitors are much more potent in the ability to inhibit HAUSP and drive N-Myc proteolysis when compared with the commercially available p5091.

In Figure 3, we propose a model whereby HAUSP inhibition can lead to tumor regression in both a p53-dependent and p53-independent context. Specifically, HAUSP inhibition drives the proteolysis of both negative regulators of p53, Mdm2 and Mdmx, which subsequently decreases the ubiquitinated-p53 forms leading to a total increase in p53 protein levels. The residual non-ubiquitinated Mdm2 can be further targeted by Nutlin-3, in combination with HAUSP inhibitors, to further activate p53 and induce apoptosis. In addition, HAUSP can regulate the proto-oncogene N-Myc in a p53-indpenedent manner (ii; Figure 3). Unlike Nutlin-3, which only activates p53, HAUSP inhibitors have the ability to both activate p53 and inhibit N-Myc leading to a dual mechanism of tumor inhibition [16]. Recently, researchers identified a novel mechanism where HAUSP inhibition augmented cancer immunotherapy by dampening T-regulator cells [26]. Along with our recent work, the therapeutic application for both p53-dependent and p53-indpendent mechanisms of HAUSP regulation may have important future implications for cancer treatments.

Figure 3.

Schematic diagram illustrating p53-dependent and -independent functions of HAUSP. (i) p53-dependent functions: HAUSP binds and deubiquitinates both Mdm2 and Mdmx. Inhibiting the deubiquitination function of HAUSP leads to Mdm2 and Mdmx ubiquitination/proteasomal degradation which subsequently stabilizes p53 protein levels to induce cell death or cell growth inhibition. In combination with the Mdm2/p53 inhibitor, Nutlin-3, the residual pool of Mdm2 can no longer ubiquitinate p53 and p53 levels strikingly increase. (ii) p53-independent function: HAUSP binds and deubiquitinates the proto-oncogene N-Myc. Inhibiting the enzymatic activity of HAUSP leads to N-Myc ubiquitination/proteasomal degradation and tumor regression in neuroblastoma.

Recently, two groups established novel, high-throughput screening techniques to identify highly-specific and highly-potent lead compound inhibiting HAUSP activity [27,28]. Unlike the commercially available inhibitors, both these novel HAUSP inhibitors (GNE-6776 [28]; and FT-671[27]) non-covalently target HAUSP. Indeed, both these specific inhibitors, to varying degrees, were able to block the deubiquitination function of MDM2 which drove p53 stabilization. Importantly, both compounds were able to induce tumor suppression in different xenograft models, albeit at high concentrations from 100–200 mg/kg. Interestingly, FT-671 was specifically demonstrated to also function in a p53-indepedent manner [27]. Further studies are needed to address the therapeutic potential of HAUSP inhibitors or in combination with other chemotherapeutic agents.

Material and methods

Western blot analysis

For western blot analysis, cells were lysed in cold Flag lysis buffer (50 mM Tris-HCl (pH 7.3), 137 mM NaCl, 10 mM NaF, 0.5 mM EDTA, 1% Triton X-100, 0.2% sarkosyl, 10% glycerol and freshly supplemented protease inhibitor cocktail (Sigma)). Nitrocellulose membranes were blocked with 5% (wt/vol) nonfat dry milk in Tris-buffered saline with Tween-20, incubated with the indicated primary antibodies followed by the HRP-conjugated secondary antibodies (GE Healthcare) (1:3,000; diluted in 2% (wt/vol) nonfat dry milk), and detected on autoradiographic films after incubating with the ECL (GE Healthcare) or SuperSignal West Dura reagents (Thermo Scientific) as explained in [16].

Antibodies

Antibodies used in this study include those specific for: N-Myc (OP-13; 1:1,000) from EMD Biosciences; p53 (DO-1; 1:1,000) and p21 (C-19; 1:500) from Santa Cruz Biotechnology; Actin (AC-15; 1:10,000), Vinculin (V284; 1:1,000) from Sigma; PARP1 (46D11; 1:1,000) from Cell Signaling; HAUSP (1:1000) from Bethyl.

Cell lines and culture

All cell lines were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS, 100 units/ml penicillin and 100 µg/ml streptomycin. Native colorectal carcinoma HCT116 and SK-N-DZ cells were described previously [16]. Doxycyclin-inducible sh-GFP or shHAUSP HCT116 p53^+/+ or HCT116 p53^−/− cells were kindly provided by S.A. Aaronson and described in [24]. Cells were verified negative for mycoplasma contamination before experiments. None of the cell lines used is reported in the International Cell Line Authentication Committee (ICLAC) database of misidentified cell lines.

HAUSP and Mdm2 inhibitors

All inhibitors were diluted in DMSO and used at indicated concentrations for each experiment. Nutlin-3 (S8059; Selleckchem). Commercially available HAUSP inhibitors: HBX41108 (4285; TOCRIS bioscience), p5091 (S7132; Selleckchem) and p22077 (662142; Calbiochem). C1 and C5 were synthesized as previously described [23].

Funding Statement

This work was supported by the National Cancer Institute [grant number CA190477], [grant number CA216884], [grant number CA085533], [grant number CA193890].

Acknowledgements

This work was supported by the National Cancer Institute of the National Institutes of Health under Award 5R01CA193890, 5RO1CA190477, 5RO1216884 and 5RO1CA085533 to W.G. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Disclosure of potential conflicts of interest

The authors declare no competing financial interests.