Acute pharmacological degradation of ERK5 does not inhibit cellular immune response or proliferation

SUMMARY

Recent interest in the role that extracellular signal-regulated kinase 5 (ERK5) plays in various diseases, particularly cancer and inflammation, has grown. Phenotypes observed from genetic knockdown or deletion of ERK5 suggested that targeting ERK5 could have therapeutic potential in various disease settings, motivating the development ATP-competitive ERK5 inhibitors. However, these inhibitors were unable to recapitulate the effects of genetic loss of ERK5, suggesting that ERK5 may have key kinase-independent roles. To investigate potential non-catalytic functions of ERK5, we report the development of INY-06-061, a potent and selective heterobifunctional degrader of ERK5. In contrast to results reported through genetic knockdown of ERK5, INY-06-061-induced ERK5 degradation did not induce anti-proliferative effects in multiple cancer cell lines or suppress inflammatory responses in primary endothelial cells. Thus, we developed and characterized a chemical tool useful for validating phenotypes reported to be associated with genetic ERK5 ablation and for guiding future ERK5-directed drug discovery efforts.

In Brief

You et al report the development and characterization of INY-06-061, a potent and highly selective ERK5 degrader. While pharmacological ERK5 degradation did not result in potent anti-proliferative or anti-inflammatory effects in this study, INY-06-061 can serve as a useful chemical probe to further dissect/investigate the biological functions of ERK5.

Graphical Abstract

INTRODUCTION

Extracellular signal-regulated kinase 5 (ERK5) is a relatively understudied member of the mitogen-activated protein kinase (MAPK) family that has been associated with a wide range of cellular functions, including proliferation and inflammation (Finegan et al., 2015; Zhou et al., 1995). Unlike the conventional MAPKs, ERK5 contains a unique, 400 amino acid C-terminal domain that consists of a nuclear localization signal (NLS) and transcriptional activation domain (TAD) (Kasler et al., 2000). Upon cellular stimulation by mitogens, agonists of Toll-like receptor-2 (TLR2) and cellular stresses, mitogen-activated protein kinase kinase 5 (MEK5) phosphorylates the TEY motif in the activation loop of ERK5, leading to ERK5 activation (Abe et al., 1996; Kato et al., 1997; Mody et al., 2003; Wilhelmsen et al., 2012). Consequently, the activated kinase domain of ERK5 auto-phosphorylates its C-terminal domain at multiple sites, promoting its own nuclear translocation (Kondoh et al., 2006; Morimoto et al., 2007). While ERK5 can directly phosphorylate various transcription factors, such as myocyte enhancer factor 2 (MEF2), the non-catalytic C-terminal tail has also been shown to interact with transcription factors and mediate gene expression (Kato et al., 1997; Sohn et al., 2005).

As a key integrator of cellular signaling, ERK5 may play important roles in various diseases, including cancer and inflammation (Wang and Tournier, 2006). For example, overexpression or constitutive activation of MEK5 or ERK5 has been observed in various malignancies, such as prostate cancer and hepatocellular carcinoma, and genetic knockdown of ERK5 resulted in suppressed proliferation of various tumor models (Hoang et al., 2017; Mehta et al., 2003; Zen et al., 2009). In addition, knockdown of ERK5 indicated that it mediates inflammation by promoting the expression of inflammatory cytokines in primary human endothelial cells and monocytes (Wilhelmsen et al., 2015).

As studies involving genetic deletion or knockdown of ERK5 suggested that ERK5 may be a promising therapeutic target, multiple ATP-competitive ERK5 inhibitors were developed. The first reported ERK5 inhibitor, XMD8-92, demonstrated promising in vitro and in vivo anti-tumor efficacy, as well as anti-inflammatory activity (Wilhelmsen et al., 2015; Yang et al., 2010). However, further investigation revealed that the biological effects of XMD8-92 and its analogue, XMD17-109, stem from off-target activities against the bromodomain and extra-terminal (BET) family of proteins, as the subsequently developed selective ERK5 inhibitors AX15836 and BAY-885 had no anti-proliferative or anti-inflammatory effects (Lin et al., 2016; Nguyen et al., 2019). This was in contrast to phenotypes observed from genetic ERK5 knockdown, suggesting that kinase-independent activities of ERK5 may play crucial roles in ERK5 signaling (Simões et al., 2016). Furthermore, some ERK5 inhibitors, such as XMD17-109 and AX15836, have been reported to drive nuclear translocation of ERK5, leading to paradoxical activation of ERK5 transcriptional activity (Lochhead et al., 2020). Thus, new chemical strategies are necessary to explore the biological functions of ERK5.

Heterobifunctional degraders, also known as PROTACs (proteolysis targeting chimeras), are molecules in which a small molecule that binds to select target proteins is chemically conjugated to a ligand of an E3 ubiquitin ligase. These bivalent molecules are able to recruit the E3 ligase into close proximity to the target protein and thereby induce its ubiquitination and subsequent proteasomal degradation (Nalawansha and Crews, 2020; Winter et al., 2015). Unlike traditional small molecule inhibitors, degraders enable acute pharmacological depletion of target proteins. Therefore, development of ERK5 degraders would allow direct chemical knockdown of ERK5 that more closely represents genetic knockdown or deletion studies, as well as investigation of the non-enzymatic functions of ERK5 that are not possible to assess with ATP-competitive inhibitors.

Here, we characterize INY-06-061, a highly selective and potent ERK5 degrader. In contrast to genetic depletion of ERK5, acute pharmacological degradation of ERK5 via treatment with INY-06-061 did not result in anti-proliferative effects in a panel of cancer cell lines. Moreover, INY-06-061-induced ERK5 degradation did not suppress inflammatory cytokine responses in primary endothelial cells. While INY-06-061 treatment did not lead to any pharmacological consequences in the biological contexts explored in this study, INY-06-061 is a useful chemical probe to interrogate the discrepancies observed between ERK5 kinase inhibition and genetic ERK5 ablation.

RESULTS

Discovery and characterization of initial degraders INY-05-091 and INY-05-128

To develop degraders of ERK5, we initially designed heterobifunctional compounds that incorporated the benzodiazepine analogues XMD17-109 and JWG-071 as ERK5 binding moieties (Figure 1A). Both XMD17-109 and JWG-071 have high binding affinities for ERK5, with reported IC₅₀ values of 162 and 88 nM, respectively. Although XMD17-109 also binds to and inhibits BRD4 and other members of the BET family proteins, JWG-071 is an optimized analogue with significantly reduced affinity for the off-target BET family proteins (Wang et al., 2018). Conjugation of a von Hippel-Landau (VHL) binding moiety (Raina et al., 2016) to XMD17-109 and JWG-071 with a six-hydrocarbon linker resulted in the bivalent compounds INY-05-091 and INY-05-128, respectively (Figure 1A).

Figure 1: Characterization of INY-05-091 and INY-05-128.

(A) Chemical structures of reported ERK5 inhibitors and initial ERK5 degraders INY-05-091 and INY-05-128. (B) Immunoblots for ERK5, BRD4, AURKA, and Actin from MOLT4 cells treated with INY-05-091 or INY-05-128 at indicated concentrations for 5 hours. Representative of 3 biological replicates. (C) MOLT4 cells were treated with the indicated compounds and concentrations for 3 days. Anti-proliferative effects of compounds were assessed using CellTiter-Glo. Error bars represent standard deviation of three biological replicates. (D) Quantitative proteomics showing relative abundance of proteins from MOLT4 cells treated for 5 hours with 250 nM of INY-05-091. (E) Quantitative proteomics showing relative abundance of proteins from MOLT4 cells treated for 5 hours with 250 nM of INY-05-128.

Linker attachment at the terminal piperazine did not compromise ERK5 binding, as both INY-05-091 and INY-05-128 retained high binding affinities for ERK5 (Table S1). Moreover, INY-05-091 and INY-05-128 induced ERK5 degradation after 5-hour treatment in MOLT4 cells (Figure 1B), with DC₅₀ values of 167 and 281 nM, respectively (Figure S1). Consistent with the selectivity profiles of the parental inhibitors XMD17-109 and JWG-071, INY-05-091 induced potent BRD4 degradation, while treatment with INY-05-128 had no effect on BRD4 protein levels (Figure 1B).

We next examined the anti-proliferative effects of the two heterobifunctional degraders in MOLT4 cells. As expected of compounds that target BRD4, both INY-05-091 and its parental inhibitor XMD17-109 displayed potent anti-proliferative activities (Figure 1C). Notably, we observed that JWG-071 and INY-05-128 also exhibited anti-proliferative activities, with EC₅₀ values of 2.4 and 1.1 μM, respectively, even though they did not target BET family proteins. In contrast, AX15836 and BAY-885, both of which are reported to be ERK5 selective inhibitors, did not have strong anti-proliferative effects, with EC₅₀s over 10 μM (Figure 1C).

Based on previous reports (Lin et al., 2016; Ramsay et al., 2011), we had hypothesized that depletion of ERK5 would have different pharmacology than inhibition of its kinase domain. To rule out the possibility that the observed anti-proliferative activity was due to off-targets, we performed a global assessment of the selectivity profiles of INY-05-091 and INY-05-128 through global, quantitative multiplexed mass spectrometry (MS)-based proteomics analysis. As expected, this analysis revealed potent downregulation of ERK5 (MAPK7) protein levels by both INY-05-091 and INY-05-128, with BRD3 and BRD4 as additional downregulated targets for INY-05-091. Surprisingly, both compounds also induced downregulation of the cell cycle regulator Aurora Kinase A (AURKA) (Figures 1D and 1E; Table S2), which was validated through immunoblot analysis (Figure 1B). As such, we attributed the potent anti-proliferative activity of INY-05-128 to the loss of AURKA signaling rather than depletion of ERK5. Therefore, while INY-05-091 and INY-05-128 demonstrate that ERK5 can be successfully targeted for degradation via VHL-based heterobifunctional degraders, the multi-targeted selectivity profiles of both compounds make them inappropriate tools for studying ERK5 biology.

Identification of INY-06-061, a potent and highly selective ERK5 degrader

To generate selective ERK5-targeting degraders, we sought to develop compounds based off of the selective ERK5 inhibitor BAY-885 (Nguyen et al., 2019). However, synthesis of BAY-885 with the pyrido[3,2-d]pyrimidine core resulted in poor yields, leading to difficulties carrying out structural-activity relationships (SAR) studies for PROTAC development. Analysis of published SAR for BAY-885 revealed that removal of the N5 nitrogen to convert the pyrido[3,2-d]pyrimidine core into a quinazoline did not significantly compromise ERK5 binding, leading to the development of INY-06-086 (Figure S2A). Not only did INY-06-086 retain high binding affinity towards ERK5 (Table S1), but its biochemical selectivity profile across the kinome was similarly selective for ERK5 as BAY-885 (Figure S2B; Table S3) (Nguyen et al., 2019).

We next synthesized an ERK5 degrader with a six-hydrocarbon linker conjugating INY-06-086 to a VHL ligand to obtain INY-06-061 (Figure 2A). INY-06-061 retained comparable binding affinity for ERK5 (K_d = 12 nM) as BAY-885 (K_d = 1 nM), demonstrating that linker attachment at the terminal piperazine did not impede the ability of INY-06-061 to bind to ERK5 (Table S1). In addition, INY-06-061 displayed a similar selectivity profile to both the parental inhibitor INY-06-086 and the original ERK5 inhibitor BAY-885 (Nguyen et al., 2019) across a panel of 468 kinases at 1 μM, with high selectivity for ERK5 (Figure S2C; Table S3).

Figure 2: Characterization of INY-06-061.

(A) Chemical structure of INY-06-061. (B) Immunoblots of ERK5, BRD4, AURKA, and Actin in MOLT4 cells treated with INY-06-061 at indicated concentrations for 5 hours. Representative of 2 biological replicates (C) Quantitative proteomics showing relative abundance of proteins in MOLT4 cells treated for 5 hours with 100 nM of INY-06-061. (D) Immunoblots of ERK5 and Actin in MOLT4 cells pre-treated with bortezomib (0.5 μM), MLN-4924 (1 μM), and BAY-885 (10 μM) for 0.5 hours, followed by treatment with INY-06-061 (100 nM) for 5 hours. Representative of 2 biological replicates (E) Immunoblots of ERK5 and Actin in MOLT4 cells treated with INY-06-061 or INY-06-089 at indicated concentrations for 5 hours. Representative of 2 biological replicates.

After verifying that INY-06-061 engaged ERK5 biochemically, we assessed the ability of the compound to destabilize ERK5 in MOLT4 cells. We found that INY-06-061 induced potent ERK5 degradation in a dose-dependent manner after 5-hour treatment in MOLT4 cells, with a DC₅₀ value of 21 nM (Figures 2B and S3A). At concentrations of 5 μM or greater, we observed diminished ERK5 degradation consistent with the hook effect, in which independent engagement of INY-06-061 to ERK5 and VHL prevents ternary complex formation (Figure 2B). In contrast to INY-05-091 and INY-05-128, INY-06-061 did not destabilize BRD4 or AURKA (Figure 2B), which was further verified through MS-based global proteomics profiling analysis. Of the ~7,700 proteins quantified, ERK5 (MAPK7) was the only protein whose abundance was significantly downregulated in MOLT4 cells treated with 100 nM of INY-06-061 for 5 hours (Figure 2C; Table S2), indicating that INY-06-061 is a potent and highly selective ERK5 degrader.

To verify the mechanism of action of INY-06-061, we pre-treated cells with either the proteasome inhibitor bortezomib or the NEDD8-activating enzyme E1 (NAE1) inhibitor MLN4924, which prevents activation of cullin-RING ligases such as CRL2^VHL. We found that this prevented ERK5 destabilization, demonstrating that INY-06-061-induced ERK5 degradation was dependent on the ubiquitin-proteasome system (Figure 2D). In addition, pre-treatment with excess quantities of BAY-885 to compete for ERK5 binding prevented INY-06-061-induced ERK5 degradation, demonstrating that engagement to ERK5 was required (Figure 2D). Finally, we synthesized INY-06-089, a negative control analogue of INY-06-061 that incorporates a diastereoisomer of the VHL ligand with significantly compromised binding affinity for VHL (Figure S3B) (Raina et al., 2016). INY-06-089 was unable to induce ERK5 downregulation at concentrations up to 5 μM, indicating that INY-06-061 acts in a proteasome- and VHL- dependent manner (Figures 2E and S3).

Pharmacological downregulation of ERK5 does not display anti-proliferative effects in various cancer models

Next, using the chemical tools developed here, as well as previously published ERK5 inhibitors, we explored the pharmacological consequences of ERK5 inhibition and degradation on cell proliferation in various cellular cancer models. As weak inhibition of potential off-targets (AXL, CDK19, SgK110, CDKL2, and AURKC) of INY-06-061 could lead to anti-proliferative effects, we compared the anti-proliferative effects of INY-06-061 to INY-06-089, as INY-06-089 also inhibits the same kinases while sparing ERK5 degradation (Figure S2B; Table S3). In cell lines reported to have constitutively active ERK5 signaling (BT-474) (Esparís-Ogando et al., 2002) or genomic amplification of ERK5 (SNU-449) (Zen et al., 2009), INY-06-061 did not induce degradation-dependent anti-proliferative effects (Figures 3A and 3B). Specifically, in both BT-474 and SNU-449 cells, the EC₅₀ values of INY-06-061 and the negative control analogue INY-06-089 were over 20 μM, similar to that of the parental ligand INY-06-086 and the reported ERK5 inhibitors AX15836 and BAY-885 (Figures 3A and 3B). As expected from its BRD4 inhibitory activity, XMD17-109 was the only compound to display anti-proliferative effects in both cell lines.

Figure 3: Pharmacological downregulation of ERK5 does not display anti-proliferative effects in various cancer models.

Anti-proliferative effects in (A) BT-474, (B) SNU-449 and (C) MM.1S cells after 3d treatment with indicated compounds and concentrations. Error bars represent standard deviation of three biological replicates. MM.1S cells were pre-treated with (D) 500 nM or (E) 1 μM of indicated compounds, followed by addition of 5 nM recombinant human IL-6. Anti-proliferative effects were assessed after 3d treatment. Error bars represent standard deviation of three biological replicates. Significance was assessed by two-way ANOVA with Bonferroni’s correction for multiple comparisons (**** indicates p ≤ 0.0001).

To broadly profile the anti-proliferative activity of INY-06-061, we submitted INY-06-061 to the PRISM platform (Broad Institute) to evaluate its anti-proliferative activity in a panel of 750 cell lines. Consistent with Dependency Map (DepMap) data (Meyers et al., 2017) which indicates a lack of dependency on ERK5 expression for cell growth, there were no cell lines identified to be sensitive to INY-06-061 treatment, with all IC₅₀ values observed to be above 1 μM (Table S4).

As the PRISM screen was carried out in cells at basal conditions, we next investigated the effects of ERK5 degradation on cell proliferation of stimulated cells. Previous studies have reported that IL-6 activates ERK5 in multiple myeloma MM.1S cells and that overexpressing a dominant negative form of ERK5 inhibits IL-6 induced cell proliferation (Carvajal-Vergara et al., 2005). Therefore, we compared the proliferation of IL-6 stimulated MM.1S cells in the presence of various ERK5 modulators at 500 nM and 1 μM. At basal conditions, INY-06-061 did not induce potent anti-proliferative effects, with an EC₅₀ of 6 μM (Figure 3C). Treatment of MM.1S cells with IL-6 promoted proliferation as expected, but treatment with INY-06-061 or the selective ERK5 inhibitors AX15836 and BAY-885 had no effect on IL-6-induced proliferation at both 500 nM and 1 μM. In contrast, the dual BRD4/ERK5 inhibitor XMD17-109 showed significant inhibition of IL-6-induced MM.1S cell proliferation (Figures 3D and 3E). Thus, our study demonstrates that acute pharmacological degradation of ERK5 has no anti-proliferative effects in multiple cancer cell lines previously reported to be regulated by ERK5.

INY-06-061-induced ERK5 degradation does not reduce pro-inflammatory cytokine secretion in human endothelial cells

Previous studies have reported that ERK5 mediates pro-inflammatory responses in endothelial cells (ECs) upon inflammatory stimulation (Wilhelmsen et al., 2015). While selective ERK5 inhibitors failed to suppress IL-6 and IL-8 secretion in human umbilical vein endothelial cells (HUVECs) upon inflammatory stimulation, siRNA-mediated knockdown of ERK5 was reported to significantly reduce the secretion of IL-6 and IL-8 (Lin et al., 2016). This contrasting phenotype between small molecule-mediated inhibition of ERK5 kinase activity and siRNA-mediated knockdown of ERK5 suggested that non-catalytic functions of ERK5 may regulate inflammatory responses in ECs.

To determine whether ERK5 degradation could recapitulate the reported effects of ERK5 knockdown on EC inflammation, we first assessed the ability of INY-06-061 to degrade ERK5 in HUVECs. 5-hour treatment with 1 μM of INY-06-061 revealed potent downregulation of ERK5, suggesting that 5-hour pre-treatment before inflammatory stimulation would be sufficient to significantly reduce ERK5 protein levels (Figure 4A). We next pre-treated HUVECs with reported ERK5 inhibitors (XMD8-92, XMD17-109, BAY-885) and the bivalent degrader INY-06-061 for 5 hours. In addition to the negative control analogue INY-06-089, we also pre-treated with a combination of BAY-885 and JQ1 to mimic the pharmacology of XMD8-92 and XMD17-109. After 5-hour pre-treatment, the cells were stimulated with lipopolysaccharide (LPS), a Toll-like receptor 4 (TLR4) agonist, after which the culture supernatants were subjected to immunoassays to measure secreted IL-6 and IL-8 levels. As expected, pre-treatment with the non-selective ERK5 inhibitors XMD8-92 and XMD17-109 (Wilhelmsen et al., 2015), as well as the combination of BAY-885 and JQ1 led to reduced IL-6 and IL-8 secretion upon LPS stimulation, while BAY-885 and the negative control analogue INY-06-089 had no effect (Figures 4B and 4C). Notably, pre-treatment with INY-06-061 also did not significantly suppress IL-6 and IL-8 secretion upon LPS stimulation (Figures 4B and 4C), in contrast to the reported genetic knockdown studies (Lin et al., 2016; Wilhelmsen et al., 2015).

Figure 4: INY-06-061-induced ERK5 degradation does not reduce pro-inflammatory cytokine secretion in human endothelial cells.

(A) Immunoblots of ERK5 and Actin from HUVECs treated with indicated compounds and concentrations for 5 hours. Representative of 2 biological replicates (B) IL-6 and (C) IL-8 secretion levels were quantified in HUVECs pre-treated with indicated compounds and concentrations for 5 hours, followed by LPS stimulation (10 μg/mL) for 6 hours. Significance was assessed by two-way ANOVA with Bonferroni’s correction for multiple comparisons (**** indicates p < 0.0001). (D) Immunoblots of ERK5 and Actin from HUVECs treated with the indicated compounds and concentrations for 24 hours. Representative of 2 biological replicates. (E) IL-6 and (F) IL-8 secretion levels were quantified in HUVECs pre-treated with indicated compounds and concentrations for 24 hours, followed by LPS stimulation (10 μg/mL) for 6 hours. Significance was assessed by two-way ANOVA with Bonferroni’s correction for multiple comparisons (* indicates p ≤ 0.05; ** indicates p ≤ 0.01; *** indicates p ≤ 0.001; **** indicates p ≤ 0.0001).

To determine whether extended ERK5 degradation could better recapitulate previously reported phenotypes, we also pre-treated HUVECs with the same set of compounds for 24 hours before stimulating with LPS. While INY-06-061-induced ERK5 degradation was maintained for 24 hours (Figure 4D), the longer pre-treatment did not affect IL-6 and IL-8 secretion levels. Consistent with the 5-hour pre-treatment, only XMD8-92, XMD17-109 and the combination of BAY-885 and JQ1 significantly reduced IL-6 and IL-8 secretion, while BAY-885 and INY-06-061 had no effect (Figures 4E and 4F). Therefore, our study indicates that INY-06-061-induced ERK5 destabilization does not suppress IL-6 and IL-8 secretion upon inflammatory stimulation.

DISCUSSION

Recent identification of ERK5 as a potential therapeutic target in cancer and inflammation prompted medicinal chemistry campaigns that resulted in the development of selective ERK5 inhibitors such as AX15836 and BAY-885 (Lin et al., 2016; Nguyen et al., 2019). However, these inhibitors were not able to recapitulate phenotypes observed through genetic modulation of ERK5 abundance or activity, such as siRNA knockdown or overexpression of a kinase dead ERK5 mutant (Lin et al., 2016). While explanations such as potential kinase-independent functions of ERK5 or paradoxical activation of ERK5-regulated transcription by ERK5 inhibitors (Lochhead et al., 2020) have been proposed, no chemical probes have been available to interrogate the discrepancies observed between ERK5 inhibition and genetic depletion or inactivation.

To develop chemical tools for acute, pharmacological destabilization of ERK5, we initially characterized the non-selective ERK5 degraders INY-05-091 and INY-05-128, which incorporated XMD17-109 and JWG-071 as parental ligands, respectively (Wang et al., 2018). Consistent with previous reports, we re-confirmed engagement of the BET family of proteins by XMD17-109, as INY-05-091 induced potent off-target degradation of BRD4. On the other hand, INY-05-128, based on the more selective inhibitor JWG-071, did not affect protein levels of the BET family of proteins. However, further characterization of JWG-071 and INY-05-128 not only revealed different anti-proliferative effects compared to selective ERK5 inhibitors, but global proteomics analysis also revealed potent AURKA downregulation by INY-05-128. Thus, phenotypes observed by JWG-071 and INY-05-128 may be caused by off-target inhibition of AURKA signaling, which highlights the importance of identifying off-target activities of small molecules.

Further medicinal chemistry efforts led to the characterization of INY-06-061, which induced potent and highly selective destabilization of ERK5 protein levels. As treatment with INY-06-061 leads to rapid pharmacological depletion of ERK5 protein, any potential scaffolding and transcriptional functions of ERK5 should be eliminated. However, INY-06-061-induced ERK5 degradation did not recapitulate the anti-proliferative or anti-inflammatory phenotypes reported through genetic knockdown of ERK5. As many of the reported studies above investigated the role of ERK5 through genetic knockdown methods with RNA interference (RNAi), the phenotypes observed may have resulted from non-specific or off-target effects of RNAi (Birmingham et al., 2006; Jackson et al., 2006). In addition, as INY-06-061 is unlikely to induce 100% degradation of the available ERK5 protein pool, residual levels of ERK5 may still be sufficient to mediate ERK5 signaling. Thus, our study suggests that non-catalytic functions of ERK5 and paradoxical activation of ERK5 transcription through ERK5 inhibition may not account for the phenotypic differences observed between ERK5 inhibition and genetic knockdown, at least in the in vitro and ex vivo settings studied here.

Limitations of the study

In addition to its effects on cellular proliferation and inflammation, genetic approaches have identified roles for ERK5 in several biological processes. For example, Cre-mediated deletion of floxed Erk5 in murine models indicated that targeting ERK5 may have therapeutic benefits by inhibiting angiogenesis in diseases dependent on neovascularization (Hayashi et al., 2004), as well as impairing protumor macrophage polarization within the tumor microenvironment (Giurisato et al., 2018). Furthermore, CRISPR-Cas9-mediated ERK5 knockout in a subset of triple negative breast cancer models displayed impaired tumor growth and metastasis in vivo, although no significant proliferative differences were observed in vitro (Hoang et al., 2020). These studies suggest that ERK5 may play complex biological roles that are difficult to recapitulate in vitro and emphasize that further investigations, especially in in vivo settings, will be required to fully understand the pharmacological consequences of ERK5 degradation. While the pharmacokinetic and pharmacodynamic properties of INY-06-061 will likely require further optimization for in vivo experiments, INY-06-061remains a useful chemical probe to validate reported phenotypes of ERK5 identified through genetic means.

Star Methods

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Nathanael S. Gray (nsgray01@stanford.edu).

Materials availability

All reported compounds generated in this manuscript will be available upon reasonable request from the lead contact with completion of an MTA.

Data and code availability

The proteomics data have been deposited in the PRIDE archive and are publicly available. Accession numbers are listed in the key resources table. No original code is reported in this paper. Any additional information required to reanalyze the data reported is available upon request from the lead contact.

Key resources table.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal anti-ERK5	Cell Signaling Technology	Cat#3372S; RRID: AB_330491
Rabbit polyclonal anti-BRD4	Fortis Life Sciences	Cat#A301-985A-M; RRID: AB_2631450
Rabbit monoclonal anti-AURKA	Cell Signaling Technology	Cat# 14475S; RRID: AB_2665504
Mouse monoclonal anti-β-Actin	Cell Signaling Technology	Cat#3700S; RRID: AB_2242334
Chemicals, peptides, and recombinant proteins
XMD17-109	MedChemExpress	Cat#HY-15665
XMD8-92	MedChemExpress	Cat#HY-14443
JWG-071	MedChemExpress	Cat#HY-108886
BAY-885	MedChemExpress	Cat#HY-112082
AX15835	MedChemExpress	Cat#HY-101846
Bortezomib	MedChemExpress	Cat#HY-10227
MLN4924	MedChemExpress	Cat#HY-70062
JQ1	MedChemExpress	Cat#HY-13030
Recombinant Human IL-6 (carrier-free)	BioLegend	Cat#570804
Lipopolysaccharides from Escherichia coli O111:B4	Millipore Sigma	Cat#L2630
INY-05-091	This study	N/A
INY-05-128	This study	N/A
INY-06-061	This study	N/A
INY-06-086	This study	N/A
INY-06-089	This study	N/A
Critical commercial assays
CellTiter Glo	Promega	Cat#G7570
LEGEND MAX™ Human IL-6 ELISA Kit	Biolegend	Cat#430507
LEGEND MAX™ Human IL-8 ELISA Kit	Biolegend	Cat#431507
ERK5 Human CMGC Kinase Binding Assay, KINOMEscan (KdELECT)	Eurofins Discovery	Cat#87-0007-1148
scanMAX Kinase Panel, KINOMEscan	Eurofins Discovery	Cat#87-0001-1000
Deposited data
INY-05-091 proteomics in MOLT4 cells	This study	PXD036104
INY-05-128 proteomics in MOLT4 cells	This study	PXD036104
INY-06-061 proteomics in MOLT4 cells	This study	PXD036105
Experimental models: Cell lines
Human: MOLT4	ATCC	Cat#CRL-1582; RRID: CVCL_0013
Human: BT-474	ATCC	Cat#HTB-20; RRID: CVCL_A4CL
Human: SNU-449	ATCC	Cat#CRL-2234; RRID: CVCL_0454
Human: MM.1S	ATCC	Cat#CRL-2974; RRID: CVCL_8792
Human: Primary Umbilical Vein Endothelial Cells; Normal, Pooled (HUVEC)	ATCC	Cat#PCS-100-013
Oligonucleotides
hERK5-HiBiT HDR Template	Integrated DNA Technologies	Cat# CD.HC9.JZMT3593
hERK5-HiBiT gRNA	Integrated DNA Technologies	Cat# Hs.HC9.DRQY5857.AA
Software and algorithms
Graphpad Prism	GraphPad Software, Inc.	https://www.graphpad.com/
Adobe Illustrator	Adobe Creative Cloud	https://www.adobe.com/creativecloud.html
Columbus image storage and analysis	PerkinElmer Informatics	https://www.perkinelmer.com/product/image-data-storage-and-analysis-system-columbus
Proteome Discoverer 2.4	Thermo Fisher Scientific	https://www.thermofisher.com/order/catalog/product/OPTON-20141?SID=srch-srp-OPTON-20141
R Framework	Team RCR: A Language and Environment for Statistical Computing	https://www.r-project.org/
Python 3.10.6	Python Software Foundation	https://www.python.org/
Other
Tandem Mass Tag (TMT) Reagents	Thermo Fisher Scientific	Cat# A44520
Bradford Reagent	BioRad	Cat# 500-0205
cOmplete, Mini Protease Inhibitor Cocktail	Sigma-Aldrich	Cat#11836153001
PhosSTOP Phosphatase Inhibitor Tablets	Sigma-Aldrich	Cat#04906837001

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell lines

MOLT4 (Male, CVCL_0013), BT-474 (Female, CVCL_A4CL), SNU-449 (Male, CVCL_0454) and MM.1S (Female, CVCL_8792) cells were cultured in RPMI (Gibco) supplemented with 10% heat inactivated fetal bovine serum and 100U/mL Penicillin-Streptomycin (Gibco) at 37 °C in the presence of 5% CO₂.

Primary Cells

Primary Umbilical Vein Endothelial Cells; Normal, Human, Pooled (HUVECs from ATCC) were grown in vascular cell basal medium (ATCC) supplemented with endothelial cell growth kit-VEGF (ATCC) and 100U/mL Penicillin-Streptomycin (Gibco) at 37 °C in the presence of 5% CO₂.

METHOD DETAILS

Immunoblotting

Cells were lysed in M-PER buffer (Thermo Scientific) containing protease and phosphatase inhibitor cocktail (Roche). Lysate concentrations were measured and normalized using a BCA assay (Pierce). Equal amounts of lysates were loaded onto 4–12% Bis-Tris gels (Invitrogen), transferred to nitrocellulose membranes (BioRad), and blocked with Intercept blocking buffer (LI-COR). Membranes were then incubated with primary antibodies against ERK5 (Cell Signaling, #3372S), AURKA (Cell Signaling, # 14475S), BRD4 (Fortis Life Sciences, # A301-985A-M), and Actin (Cell Signaling, # 3700S) overnight at 4 °C, followed by incubation with IRDye®800-labeled goat anti-rabbit IgG and IRDye®680-labeled goat anti-mouse IgG (LI-COR) secondary antibodies for detection on an Odyssey CL_x System.

In vitro kinase assays

Biochemical binding constants (K_d) of compounds to ERK5 was determined through the KdELECT assays provided by Eurofins Discovery using an 11-point dose response curve.

Biochemical selectivity assay

Biochemical selectivity across 468 kinases was measured through the scanMAX kinase assay panel provided through Eurofins Discovery.

Proliferation assays

Cell lines were plated at densities ranging from 500 to 1000 cells per well in 384-well plates. Cells were treated at the indicated concentrations for 72 hours, after which anti-proliferative effects of compounds were assessed using CellTiter-Glo (Promega). EC₅₀ values were calculated using the GraphPad Prism nonlinear regression curve fit.

For MM.1S cells stimulated with IL-6, cells were treated in 96 well plates at 10,000 cells per well. Cells were pre-treated with compounds at indicated concentrations for 5 hours before adding recombinant human IL-6 (Biolegend) at 5 nM. After 72 hours, the anti-proliferative effects were assessed using CellTiter-Glo (Promega).

IL-6 and IL-8 ELISA

HUVECs were plated at cell densities ranging from 30,000 to 60,000 cells per well in 96 well plates. The cells were pre-treated with indicated compounds for 5 or 24 hours, after which the cells were stimulated with LPS (Sigma-Aldrich) at 10 μg/mL for 6 hours. IL-6 and IL-8 levels in the supernatant were measured using LEGEND MAX™ Human IL-6 (Biolegend # 430507) and Human IL-8 (Biolegend # 431507) ELISA kits, following the manufacturer’s protocols. The data was analyzed with GraphPad Prism.

Generation of ERK5 HiBiT cell lines

Introduction of a HiBiT tag into the endogenous ERK5 locus in Molt4 cells was performed via CRISPR-Cas9 editing. ALT-R CRISPR RNA (crRNA) and trans-activating CRISPR RNA (tracrRNA) (Integrated DNA Technologies, IDT) were resuspended in nuclease-free duplex buffer at a final concentration of 100 μM. Equal volumes of crRNA and tracrRNA were mixed and heated for 5 min at 95 °C. After heating, the complex was gradually cooled to room temperature. The oligo complex was then incubated at room temperature for 20 min with Cas9 Nuclease V3 (IDT) to form the ribonucleoprotein complex. Subsequently, the double-stranded DNA HDR template (Table S5), which incorporated the HiBiT sequence into the C-terminus of the ERK5 genome, the RNP complex, and an electroporation enhancer (IDT) were co-electroporated into Molt4 cells using the Neon Electroporator (Thermo Fisher). Cells were seeded into media with HDR enhancer (IDT). Subsequently, single cells were isolated via FACS sorting, and HiBiT expression from individual clones was detected using the Nano-Glo® HiBiT Lytic Detection System (Promega). Correct insertion of the HiBiT tag in the genome of the knocked-in cells was confirmed by sequencing.

TMT LC-MS3 mass spectrometry

MOLT4 cells were treated with DMSO (biological triplicate) or INY-06-061 (100 nM), INY-05-091 (250 nM) or INY-05-128-01 (250 nM) degrader for 5 h and cells were harvested by centrifugation. Cell lysis and Tandem Mass Tagged (TMT) tryptic peptides were prepared for LC-MS analysis following procedures published (Donovan et al., 2018).

Data were collected using an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA) coupled with a Proxeon EASY-nLC 1200 LC pump (Thermo Fisher Scientific). Peptides were separated on a 50 cm 75 μm inner diameter EasySpray ES903 microcapillary column (Thermo Fisher Scientific). Peptides were separated using a 190 min gradient of 6 – 27% acetonitrile in 1.0% formic acid with a flow rate of 300 nL/min.

Each analysis used a MS3-based TMT method as described previously (McAlister et al., 2014). The data were acquired using a mass range of m/z 340 – 1350, resolution 120,000, AGC target 5 × 105, maximum injection time 100 ms, dynamic exclusion of 120 s for the peptide measurements in the Orbitrap. Data dependent MS2 spectra were acquired in the ion trap with a normalized collision energy (NCE) set at 35%, AGC target set to 1.8 × 104 and a maximum injection time of 120 ms. MS3 scans were acquired in the Orbitrap with HCD collision energy set to 55%, AGC target set to 2 × 105, maximum injection time of 150 ms, resolution at 50,000 and with a maximum synchronous precursor selection (SPS) precursor set to 10.

Proteome Discoverer 2.4 (Thermo Fisher Scientific) was used for .RAW file processing and controlling peptide and protein level false discovery rates, assembling proteins from peptides, and protein quantification from peptides. MS/MS spectra were searched against a Swissprot human database (February 2020) with both the forward and reverse sequences as well as known contaminants such as human keratins. Database search criteria were as follows: tryptic with two missed cleavages, a precursor mass tolerance of 20 ppm, fragment ion mass tolerance of 0.6 Da, static alkylation of cysteine (57.0215 Da), static TMT labeling of lysine residues and N-termini of peptides (304.2071 Da), and variable oxidation of methionine (15.9949 Da). TMT reporter ion intensities were measured using a 0.003 Da window around the theoretical m/z for each reporter ion in the MS3 scan. The peptide spectral matches with poor quality MS3 spectra were excluded from quantitation (summed signal-to-noise across channels < 100 and precursor isolation specificity < 0.5), and the resulting data was filtered to only include proteins with a minimum of 2 unique peptides quantified. Reporter ion intensities were normalized and scaled using in-house scripts in the R framework (Team, 2014).Statistical analysis was carried out using the limma package within the R framework (Ritchie et al., 2015).

General Chemistry Methods

Reagents and solvents were purchased from commercial vendors and used without further purification otherwise noted. Reactions were monitored using a Waters Acquity UPLC/MS system (Waters PDA eλ Detector, QDa Detector, Sample manager - FL, Binary Solvent Manager) using Acquity UPLC® BEH C18 column (2.1 × 50 mm, 1.7 μm particle size): solvent gradient = 85% A at 0 min, 1% A at 1.7 min; solvent A = 0.1% formic acid in Water; solvent B = 0.1% formic acid in Acetonitrile; flow rate: 0.6 mL/min. Products were purified by CombiFlash®Rf with Teledyne Isco RediSep® normal-phase silica flash columns (4 g, 12 g, 24 g, 40 g) and preparative HPLC using Waters SunFireTM Prep C18 column (19 × 100 mm, 5 μm particle size) using a gradient of 0–100% methanol in water containing 0.05% trifluoroacetic acid (TFA) over 48 minutes at a flow of 40 mL/min. ¹H NMR spectra were recorded on 500 MHz Bruker Avance III spectrometer, and chemical shifts are reported in million (ppm, δ) downfield from tetramethylsilane (TMS). Coupling constants (J) are reported in Hz. Spin multiplicities are described as s (singlet), br (broad singlet), d (doublet), t (triplet), q (quartet) and m (multiplet). Purities of assayed compounds were in all cases greater than 95%, as determined by reverse-phase HPLC analysis.

Synthesis of INY-05-091 and INY-05-128

tert-butyl-6-(4-(4-((11-cyclopentyl-5-methyl-6-oxo-6,11-dihydro-5H-benzo[e]pyrimido[5,4-b][1,4]diazepin-2-yl)amino)benzoyl)piperazin-1-yl)hexanoate (2a):

To 1a (20 mg, 0.033 mmol) dissolved in MeCN (1 mL) was added potassium carbonate (18 mg, 0.132 mmol) and tert-butyl 6-bromohexanoate (12 mg, 0.049 mmol), which was stirred at 80 °C overnight. Next day, the reaction mixture was diluted with ethyl acetate (15 mL) and washed with brine (4 × 5mL). The organic layer was dried over anhydrous sodium sulfate, filtered, and concentrated in vacuo. The crude material was purified by column chromatography on silica gel (0–20% MeOH/DCM) to afford 2a (18 mg, 82% yield). LC-MS: m/z 668.51 [M+H]⁺.

6-(4-(4-((11-cyclopentyl-5-methyl-6-oxo-6,11-dihydro-5H-benzo[e]pyrimido[5,4-b][1,4]diazepin-2-yl)amino)benzoyl)piperazin-1-yl)hexanoic acid (3a):

To 2a (18 mg, 0.027 mmol) was added 800 μL of DCM and 200 μL of TFA. The reaction was stirred for 4 hours at room temperature. The reaction mixture was concentrated in vacuo to afford crude 3a (quantitative yield), which was used directly for the next reaction. LC-MS: m/z 612.49 [M+H]⁺.

(2S,4R)-1-((S)-2-(6-(4-(4-((11-cyclopentyl-5-methyl-6-oxo-6,11-dihydro-5H-benzo[e]pyrimido[5,4-b][1,4]diazepin-2-yl)amino)benzoyl)piperazin-1-yl)hexanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-((S)-1-(4-(4-methylthiazol-5-yl)phenyl)ethyl)pyrrolidine-2-carboxamide (INY-05-091):

To a solution of 3a (17 mg, 0.027 mmol) and (2S,4R)-1-((S)-2-amino-3,3-dimethylbutanoyl)-4-hydroxy-N-((S)-1-(4-(4-methylthiazol-5-yl)phenyl)ethyl)pyrrolidine-2-carboxamide (13 mg, 0.027 mmol) in DMF (1 mL) was added HATU (10 mg, 0.027 mmol) and DIEA (28 μL, 0.16 mmol). The reaction was stirred for 1 hour. The reaction mixture was purified by reverse-phase HPLC (30–90% methanol in water) to obtain INY-05-091 as a TFA salt (10 mg, 29% yield). LC-MS: m/z 1038.62 [M+H]⁺. ¹H NMR (500 MHz, DMSO) δ 9.95 (s, 1H), 9.83 (s, 1H), 9.00 (s, 1H), 8.49 (s, 1H), 8.36 (d, J = 7.8 Hz, 1H), 7.90 – 7.78 (m, 3H), 7.59 (dd, J = 7.8, 1.7 Hz, 1H), 7.50 – 7.37 (m, 6H), 7.36 – 7.27 (m, 1H), 7.18 (td, J = 7.5, 1.0 Hz, 1H), 4.96 – 4.87 (m, 1H), 4.73 (p, J = 6.4 Hz, 1H), 4.54 (d, J = 9.4 Hz, 1H), 4.42 (t, J = 8.1 Hz, 1H), 4.32 – 4.27 (m, 1H), 3.66 – 3.56 (m, 2H), 3.56 – 3.48 (m, 2H), 3.45 (s, 3H), 3.09 (s, 4H), 2.46 (s, 3H), 2.39 – 2.25 (m, 2H), 2.20 – 2.09 (m, 2H), 2.06 – 1.99 (m, 1H), 1.84 – 1.78 (m, 1H), 1.74 – 1.40 (m, 11H), 1.38 (d, J = 7.0 Hz, 3H), 1.34 – 1.20 (m, 4H), 0.94 (s, 9H).

(2S,4R)-1-((2S)-2-(6-(4-(4-((11-(sec-butyl)-5-methyl-6-oxo-6,11-dihydro-5H-benzo[e]pyrimido[5,4-b][1,4]diazepin-2-yl)amino)benzoyl)piperazin-1-yl)hexanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-((S)-1-(4-(4-methylthiazol-5-yl)phenyl)ethyl)pyrrolidine-2-carboxamide (INY-05-128)

INY-05-128 was synthesized with similar procedure as INY-05-091 using intermediate 1b as the starting material (15 mg, 20% yield). LC-MS: m/z 1026.74 [M+H]⁺. ¹H NMR (500 MHz, DMSO) δ 9.90 (d, J = 35.8 Hz, 1H), 9.77 (s, 1H), 9.00 (s, 1H), 8.50 (d, J = 5.1 Hz, 1H), 8.36 (d, J = 7.8 Hz, 1H), 7.84 – 7.81 (m, 2H), 7.60 (t, J = 7.3 Hz, 1H), 7.49 – 7.47 (m, 1H), 7.45 – 7.43 (m, 3H), 7.40 – 7.37 (m, 2H), 7.33 – 7.27 (m, 1H), 7.19 (q, J = 6.9 Hz, 1H), 4.92 (p, J = 7.2 Hz, 1H), 4.54 (d, J = 9.3 Hz, 1H), 4.42 (t, J = 8.1 Hz, 1H), 4.34 – 4.27 (m, 2H), 4.19 – 4.14 (m, 1H), 3.66 – 3.58 (m, 2H), 3.56 – 3.48 (m, 2H), 3.44 (s, 3H),f 3.10 (s, 5H), 2.46 (s, 3H), 2.33 – 2.26 (m, 1H), 2.20 – 2.13 (m, 1H), 2.05 – 2.00 (m, 1H), 1.87 – 1.77 (m, 3H), 1.70 – 1.60 (m, 3H), 1.60 – 1.45 (m, 5H), 1.39 – 1.35 (m, 4H), 1.32 – 1.24 (m, 3H), 0.94 (s, 9H), 0.84 (t, J = 7.4 Hz, 3H).

Synthesis of INY-06-086

tert-butyl 4-(7-(4-methylpiperazin-1-yl)quinazolin-4-yl)piperidine-1-carboxylate (5)

To a solution of intermediate 4 (100 mg, 0.29 mmol) and 1-methylpiperazine (43 mg, 0.43 mmol) in toluene (3 mL) was added Pd(OAc)₂ (7 mg, 0.03 mmol), BINAP (36 mg, 0.06 mmol) and cesium carbonate (284 mg, 0.87 mmol). The reaction mixture was purged with nitrogen, and stirred overnight at 100 °C. The next day, the reaction mixture was diluted with DCM (20 mL) and filtered through celite. The organic layer was washed with sat. aq sodium bicarbonate (20 mL) and extracted with DCM (2 × 10 mL). The organic layers were combined, dried over anhydrous sodium sulfate, filtered, and concentrated in vacuo. The crude material was purified by column chromatography on silica gel (0–30% MeOH/DCM) to obtain intermediate 5 (115 mg, 97% yield). LC-MS: m/z 412.3 [M+H]⁺.

7-(4-methylpiperazin-1-yl)-4-(piperidin-4-yl)quinazoline (6)

To intermediate 5 (115 mg, 0.28 mmol) was added DCM (1.5 mL) and TFA (0.5 mL). The reaction was stirred at room temperature for 1 hour. The reaction mixture was concentrated in vacuo to obtain crude 6 (quantitative yield), which was used directly for the next reaction. LC-MS: m/z 312.25 [M+H]⁺.

(2-amino-4-(trifluoromethoxy)phenyl)(4-(7-(4-methylpiperazin-1-yl)quinazolin-4-yl)piperidin-1-yl)methanone (INY-06-086)

To a solution of intermediate 6 (87 mg, 0.28 mmol) and 2-amino-4-(trifluoromethoxy)benzoic acid (62 mg, 0.28 mmol) in DMF (2 mL) was added HATU (106 mg, 0.28 mmol) and DIEA (490 μL, 2.8 mmol). The reaction was stirred at room temperature for 1 hour, then purified by reverse-phase HPLC (15–75% methanol in water) to obtain INY-06-086 (29 mg, 14% yield). LC-MS: m/z 515.27 [M+H]⁺. ¹H NMR (500 MHz, DMSO) δ 9.96 (s, 1H), 9.03 (s, 1H), 8.33 (d, J = 9.5 Hz, 1H), 7.60 (dd, J = 9.5, 2.7 Hz, 1H), 7.28 (d, J = 2.6 Hz, 1H), 7.14 (d, J = 8.3 Hz, 1H), 6.70 – 6.65 (m, 1H), 6.54 – 6.48 (m, 1H), 4.28 (d, J = 13.1 Hz, 2H), 3.98 – 3.88 (m, 1H), 3.57 (d, J = 11.3 Hz, 2H), 3.31 – 3.12 (m, 6H), 2.88 (s, 3H), 1.94 – 1.73 (m, 4H), 1.32 – 1.21 (m, 1H).

Synthesis of INY-06-061

benzyl 4-(4-(1-(tert-butoxycarbonyl)piperidin-4-yl)quinazolin-7-yl)piperazine-1-carboxylate (7)

To a solution of intermediate 4 (389 mg, 1.12 mmol) and benzyl piperazine-1-carboxylate (370 mg, 1.68 mmol) in toluene (12 mL) was added Pd(OAc)₂ (25 mg, 0.11 mmol), BINAP (140 mg, 0.22 mmol) and cesium carbonate (1.1 g, 3.36 mmol). The reaction mixture was purged with nitrogen, and stirred overnight at 100 °C. The next day, the reaction mixture was diluted with DCM (30 mL) and filtered through celite. The organic layer was washed with sat. aq sodium bicarbonate (30 mL) and extracted with DCM (2 × 20 mL). The organic layers were combined, dried over anhydrous sodium sulfate, filtered, and concentrated in vacuo. The crude material was purified by column chromatography on silica gel (0–100% EA/DCM) to obtain 7 (507 mg, 85% yield). LC-MS: m/z 532.35 [M+H]⁺.

benzyl 4-(4-(1-(tert-butoxycarbonyl)piperidin-4-yl)quinazolin-7-yl)piperazine-1-carboxylate (8)

To 7 (507 mg, 0.95 mmol) was added DCM (6 mL) and TFA (2 mL). The reaction was stirred at room temperature for 1 hour and concentrated in vacuo to obtain crude 8 (quantitative yield). LC-MS: m/z 432.27 [M+H]⁺.

(2-amino-4-(trifluoromethoxy)phenyl)(4-(7-(piperazin-1-yl)quinazolin-4-yl)piperidin-1-yl)methanone (9)

To a solution of 8 (390 mg, 0.9 mmol) and 2-amino-4-(trifluoromethoxy)benzoic acid (199 mg, 0.9 mmol) in DMF (10 mL) was added HATU (342 mg, 0.9 mmol) and DIEA (1.6 mL, 9 mmol). The reaction was stirred at room temperature for 1 hour. The reaction mixture was diluted with ethyl acetate (50 mL) and washed with brine (5 × 10 mL). The organic layer was dried over anhydrous sodium sulfate, filtered, and concentrated in vacuo to obtain crude benzyl 4-(4-(1-(2-amino-4-(trifluoromethoxy)benzoyl)piperidin-4-yl)quinazolin-7-yl)piperazine-1-carboxylate (quantitative yield). LC-MS: m/z 635.32 [M+H]⁺. The crude material was dissolved in MeOH (10 mL)and Pd/C (50 mg) was added. H₂ (g) was introduced to the reaction, and the reaction was stirred at room temperature for 4 hours. The reaction mixture was filtered over celite, concentrated in vacuo and purified by column chromatography on silica gel (0–30% MeOH with 1.75N NH₃/DMC) to obtain 9 (166 mg, 43% yield over 2 steps). LC-MS: m/z 501.23 [M+H]⁺.

6-(4-(4-(1-(2-amino-4-(trifluoromethoxy)benzoyl)piperidin-4-yl)quinazolin-7-yl)piperazin-1-yl)hexanoic acid (10)

To 9 (27 mg, 0.054 mmol) in MeCN (1 mL) was added tert-butyl 6-bromohexanoate (27 mg, 0.11 mmol) and potassium carbonate (30 mg, 0.22 mmol). The reaction was stirred at 80 °C overnight. Next day, the reaction mixture was concentrated in vacuo and purified by column chromatography on silica gel (0–20% MeOH/DCM) to obtain tert-butyl 6-(4-(4-(1-(2-amino-4-(trifluoromethoxy)benzoyl)piperidin-4-yl)quinazolin-7-yl)piperazin-1-yl)hexanoate (36 mg, 100% yield). LC-MS: m/z 671.45 [M+H]⁺. The product was then dissolved in DCM (750 μL) and TFA (250 μL) at room temperature for 1 hour and concentrated in vacuo. The crude material 10 (quantitative yield) was directly used for the next step without further purification. LC-MC: m/z 615.33 [M+H]⁺.

(2S,4R)-1-((S)-2-(6-(4-(4-(1-(2-amino-4-(trifluoromethoxy)benzoyl)piperidin-4-yl)quinazolin-7-yl)piperazin-1-yl)hexanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-((S)-1-(4-(4-methylthiazol-5-yl)phenyl)ethyl)pyrrolidine-2-carboxamide (INY-06-061)

To 10 (33 mg, 0.054 mmol) and (2S,4R)-1-((S)-2-amino-3,3-dimethylbutanoyl)-4-hydroxy-N-((S)-1-(4-(4-methylthiazol-5-yl)phenyl)ethyl)pyrrolidine-2-carboxamide (25 mg, 0.054 mmol) in DMF (1 mL) was added HATU (21 mg, 0.054 mmol) and DIEA (47 μL, 0.27 mmol). The reaction was stirred at room temperature for 1 hour and purified by reverse-phase HPLC (35–95% methanol in water) to obtain INY-06-061 (18 mg, 26% yield). LC-MS: m/z 1041.69 [M+H]⁺. ¹H NMR (500 MHz, DMSO) δ 9.78 (s, 1H), 9.04 (s, 1H), 9.00 (s, 1H), 8.40 – 8.32 (m, 2H), 7.83 (d, J = 9.3 Hz, 1H), 7.62 (dd, J = 9.5, 2.5 Hz, 1H), 7.48 – 7.43 (m, 2H), 7.41 – 7.37 (m, 2H), 7.30 – 7.27 (m, 1H), 7.14 (d, J = 8.4 Hz, 1H), 6.70 – 6.66 (m, 1H), 6.51 (d, J = 8.4 Hz, 1H), 4.92 (p, J = 7.1 Hz, 1H), 4.54 (d, J = 9.4 Hz, 1H), 4.42 (t, J = 8.1 Hz, 1H), 4.33 – 4.23 (m, 3H), 3.99 – 3.91 (m, 1H), 3.68 – 3.55 (m, 4H), 3.33 – 3.24 (m, 2H), 3.17 (s, 6H), 2.46 (s, 3H), 2.35 – 2.27 (m, 1H), 2.21 – 2.14 (m, 1H), 2.06 – 1.99 (m, 1H), 1.92 – 1.76 (m, 5H), 1.76 – 1.66 (m, 2H), 1.62 – 1.48 (m, 3H), 1.38 (d, J = 7.0 Hz, 3H), 1.36 – 1.23 (m, 3H), 0.94 (s, 9H).

(2R,4S)-1-((S)-2-(6-(4-(4-(1-(2-amino-4-(trifluoromethoxy)benzoyl)piperidin-4-yl)quinazolin-7-yl)piperazin-1-yl)hexanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-((S)-1-(4-(4-methylthiazol-5-yl)phenyl)ethyl)pyrrolidine-2-carboxamide (INY-06-089)

INY-06-089 was synthesized with similar procedure as INY-06-061 using intermediate 10 and (2R,4S)-1-((S)-2-amino-3,3-dimethylbutanoyl)-4-hydroxy-N-((S)-1-(4-(4-methylthiazol-5-yl)phenyl)ethyl)pyrrolidine-2-carboxamide as the starting materials (33 mg, 65% yield). LC-MS: m/z 1041.67 [M+H]⁺. ¹H NMR (500 MHz, DMSO) δ 9.71 (s, 1H), 9.04 (s, 1H), 9.00 (s, 1H), 8.34 (d, J = 9.5 Hz, 1H), 8.18 (d, J = 7.9 Hz, 1H), 7.92 (d, J = 8.3 Hz, 1H), 7.61 (dd, J = 9.5, 2.5 Hz, 1H), 7.51 – 7.41 (m, 4H), 7.30 – 7.25 (m, 1H), 7.14 (d, J = 8.4 Hz, 1H), 6.69 – 6.66 (m, 1H), 6.54 – 6.48 (m, 1H), 4.90 (p, J = 6.7 Hz, 1H), 4.47 (d, J = 8.3 Hz, 1H), 4.45 – 4.39 (m, 1H), 4.35 – 4.31 (m, 1H), 4.25 (d, J = 13.2 Hz, 2H), 3.98 – 3.90 (m, 1H), 3.80 – 3.74 (m, 1H), 3.63 – 3.55 (m, 2H), 3.52 (dd, J = 10.5, 3.9 Hz, 1H), 3.25 (t, J = 12.5 Hz, 3H), 3.19 – 3.05 (m, 5H), 2.47 (s, 3H), 2.32 – 2.24 (m, 1H), 2.19 – 2.11 (m, 1H), 2.07 – 1.90 (m, 3H), 1.90 – 1.80 (m, 4H), 1.67 – 1.59 (m, 2H), 1.55 – 1.48 (m, 2H), 1.33 (d, J = 7.0 Hz, 3H), 1.28 – 1.20 (m, 2H), 0.97 (s, 9H).

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical details of experiments can be found in the figure legends or method details. GraphPad Prism 9 was used for statistical analyses.

Supplementary Material

3

Table S1: Biochemical K_d values (nM) of ERK5 binding compounds. Related to Figures 1 and 2.

4

Table S2: Proteomics output from limma processing of MOLT4 cells treated for 5 hours with INY-05-091, INY-05-128, and INY-06-061 vs DMSO control. Related to Figures 1 and 2.

5

Table S3: KINOMEscan hits of INY-06-061 and INY-06-089 (1 μM). Related to Figure 2 and Figure S2.

6

Table S4: Log₂(IC₅₀) values of INY-06-061 from the Broad PRISM Screen. Blank cells represent a lack of potency of INY-06-061 at tested concentrations (highest concentration = 10 μM) or a non-linear regression curve could not be fit to obtain Log₂(IC₅₀) values. Related to Figure 3.

7

Table S5: HDR and gRNA set designs for MOLT4 ERK5-HiBiT cells. Related to STAR methods.

Scheme 1.

Scheme 2.

Scheme 3.

Highlights.

• PROTACs that incorporate XMD17-109 and JWG-071 are not selective for ERK5

• INY-06-061 is a potent and highly selective heterobifunctional degrader of ERK5

• INY-06-061 treatment does not compromise viability in multiple cancer cell lines

• INY-06-061 does not affect pro-inflammatory cytokine secretion in endothelial cells

SIGNIFICANCE.

ERK5 function has been investigated through a wide variety of methods, including inhibition by small molecules, depletion via RNA interference and genetic knockout. As discrepancies between genetic knockdown and inhibition of ERK5 have been reported, small molecule approaches to interrogate the differences between loss of ERK5 protein and inhibition of its kinase activity are necessary. Here, we develop and characterize INY-06-061, a potent and highly selective small molecule degrader of ERK5. While our studies indicate that INY-06-061-induced ERK5 destabilization does not have potent anti-proliferative or anti-inflammatory effects in multiple cancer cell lines or HUVECs, respectively, further studies are necessary to understand the biological roles ERK5. Thus, INY-06-061 will be a valuable chemical probe to not only study the pharmacological effects of ERK5 degradation, but also to investigate the differences in phenotypes induced by ERK5 inhibition and genetic depletion.