Spirocyclic Chromenopyrazole Inhibitors Disrupting the Interaction between the RNA‐Binding Protein LIN28 and Let‐7

Small‐molecule inhibitors of the RNA‐binding and regulating protein LIN28 have the potential to be developed as chemical probes for biological perturbation and as therapeutic candidates. Reported small molecules disrupting the interaction between LIN28 and let‐7 miRNA suffer from moderate to weak inhibitory activity and flat structure‐activity relationship, which hindered the development of next‐generation LIN28 inhibitors that warrant further evaluations. We report herein the identification of new LIN28 inhibitors utilizing a spirocyclization strategy based on a chromenopyrazole scaffold. Representative compounds 2–5 showed potent in vitro inhibitory activity against LIN28‐let‐7 interaction and single‐digit micromolar potency in inhibiting the proliferation of LIN28‐expressing JAR cancer cells. The spirocyclic compound 5 incorporated a position that is amenable for functional group appendage and further structural modifications. The binding mode of compound 5 with the LIN28 cold shock domain was rationalized via a molecular docking analysis.


Introduction
RNA-binding proteins (RBPs) play vital regulatory roles in all aspects of RNA biology including biogenesis, function, and metabolism, and are thus emerging targets for the development of small-molecule modulators as chemical probes and therapeutic candidates. RBPs are understudied protein groups in comparison with other well-established targets, and only a limited number of RBPs have been probed with small-molecule modulators. [1] In parallel with the increasing interest in targeting RBPs, an increasing number of small-molecule binders of structured RNAs are being identified. [2] The interaction between LIN28 and let-7 miRNAs is one of the most extensively studied protein-RNA interactions, for which small-molecule RBP modulators have been reported. [1b] LIN28 has two isoforms in mammalian cells, LIN28A and LIN28B. Both contain an N-terminal cold shock domain (CSD) and a Cterminal zinc knuckle domain (ZKD). [3] LIN28A and LIN28B are located predominantly in cytoplasm and nucleus, respectively, suppressing the biogenesis of let-7 miRNA family. Matured let-7 negatively regulates the mRNAs of multiple oncogenes, such as RAS, MYC and HMGA2. Binding of LIN28B to let-7 blocks Drosha mediated processing of primary let-7 (pri-let-7) in the nucleus while LIN28A binding not only inhibits Dicer-mediated processing of precursor let-7 (pre-let-7) but also recruits terminal uridylyl transferase 4 (TUT4) to initiate degradation of pre-let-7 miRNAs in the cytoplasm ( Figure 1A). [4] Overexpression of LIN28A or LIN28B and downregulation of let-7 miRNAs have been observed in many human cancers and are associated with poor clinical prognosis. [5] Additionally, LIN28 has been recently identified as an epigenetic regulator that negatively influences DNA methylation by recruiting the methylcytosine dioxygenase TET1. [4a] Furthermore, it was found that LIN28 influences histone methylation in mouse pluripotent stem cells via regulation of Sadenosyl-methionine metabolism. [6] Taken together, disruption of the LIN28-pre-/pri-let-7 interaction holds great potential to be employed as a novel anticancer strategy.
Small-molecule inhibitors of different structures and activities have been reported to disrupt the LIN28-pre-/pri-let-7 interaction, [7] such as the chromenopyrazole inhibitor SB1301 (1) [7a] and the trisubstituted pyrrolinones. [7g] However, the majority of these small-molecule inhibitors suffer from limited potency and efficacy in cells. The fact that even the most potent LIN28 inhibitors reported to date showed only micromolar potency in either biochemical or cell-based assays reflects the challenging task to efficiently inhibit LIN28. This is probably because that the interaction of LIN28 and let-7 involves an extensive network of surface interactions that is comparable to that of protein-protein interactions and the presence of only surface dents or shallow pockets on LIN28 make it difficult for small molecules to engage in efficient binding and inhibition. [8] Therefore, the development of LIN28 inhibitors with new structures, mechanism of inhibition, and improved potency is a highly sought-after topic. One such promising strategy will be the development of bifunctional inhibitors by attaching an additional warhead to the LIN28-inhibiting scaffold.
Herein, we report a spirocyclization strategy to build new LIN28 inhibitors based on the chromenopyrazole scaffold of compound 1, which was one of the earliest reported CSDtargeting LIN28 inhibitors. [7a] We propose to utilize the spirocyclization strategy, because first, spirocyclization will induce a high content of sp 3 -hybridrized carbon atoms and molecular complexity and has the potential to optimize potency, selectivity and absorption, distribution, metabolism and excretion (ADME) properties of the resulting spirocyclic scaffold due to intensified rigidity. [9] Second, introduction of a rigid spirocyclic scaffold, which locks functional groups in favored orientations, [9a] poses a promising strategy to target the dynamic LIN28-pri-/pre-let-7 interaction. In this work, we evaluated spirocyclic compounds 2-5 as LIN28 CSD inhibitors with single-digit micromolar potency ( Figure 1B and Scheme 1). Furthermore, we identified the spirocyclic moiety as an amenable position for further functional group attachment, which will be crucial for the development of next-generation bifunctional LIN28 inhibitors as no such position has been identified for any other reported LIN28-targeting small molecules.

Results and Discussion
We performed extensive structural modifications by synthesizing analogues of compound 1 based on the chromenopyrazole scaffold with the aim to identify a position that tolerates structural changes without compromising LIN28 inhibitory activity. Modifications performed at R 1 -R 3 all led to reduced inhibitory activity in different extents, especially for modifications involving the replacement of the carboxylic acid at 1phenyl group on the chromenopyrazole scaffold. [13] Therefore, we switched our focus to modify the gem-dimethyl moiety in this study via a spirocyclization strategy (Scheme 1 and Table 1).
We introduced the spirocyclic moiety at the 4-position by firstly connecting the gem-dimethyl group to form fourmembered (2), five-membered (3) and six-membered (4) spirocyclic rings ( Figure 1C). To explore the possibility of appendage of an additional group, we replaced the cyclohexane moiety of 4 with a N-Bn-piperidine moiety in compound 5. The synthetic route to obtain the spirocyclic chromenopyrazoles (2-5) is shown in Scheme 1. The synthesis started with a nitrification to yield 4-fluoro-2-hydroxyacetophenone 6, followed by Cbz-piperazine substitution to give intermediate 7. The spirocyclic compounds 8-11 were then constructed via the Kabbe condensation using 7 and the corresponding cyclic aldehydes. The acetal intermediates 12-15 were prepared by using the triethyl orthoformate and converted to the aldehyde intermediates 16-19 in the presence of catalytic iodine. In the final step, cyclization with the 4hydrazinobenzoic acid gave the spirocyclic compounds 2-5.
The LIN28 inhibitory activities of the spirocyclic compounds 2-4 were first evaluated in EMSA (Table 1 and Figure 2). Compounds 2-4 showed IC 50 between 15 and 5 μM, with the spiro cyclohexane 4 showing the best inhibitory potency. The additional benzyl group in compound 5 retained the potent activity (IC 50 : 5.4 μM, Figure 2). A nano differential scanning fluorimetry (NanoDSF) measurement was performed, which did not significantly show the difference in LIN28 stabilizing effect upon binding with either compound 4 or 5 in comparison with that of compound 1, as measured by the increased melting temperature (T m ) by~2°C (compound 4) and 1°C (compound 5), respectively (Table 1). Then, a direct binding assay employing biolayer interferometry (BLI) revealed stronger binding of compound 5 to the LIN28 CSD in comparison with that of compound 1 ( Figure S1).
To study the inhibitory mode of the spirocyclic compounds, we performed molecular docking analysis based on the resolved structure of the LIN28-preE-let-7f-1 complex (  ure 3). [4c] As compound 1 was characterized as a LIN28 inhibitor binding to the CSD, [7a] different surface dents and pockets of the CSD were probed in docking analysis binding with compound 5. As shown in Figure 3A and 3B, the optimal binding mode was found at the 5 U-6A-14U-15 A binding site of the LIN28 CSD. Key interactions with LIN28 were formed in two respects as shown in Figure 3C and 3D. The 4-Cbz-piperazin-1-yl moiety is located at a hydrophobic area. Notably, the terminal benzene ring formed a π-π interaction with Arg122. Second, hydrogen bonds were predicted to form involving one that is between Arg50 and the carbonyl acid moiety, one between the nitro group and Val49, and one between the nitrogen of the N-Bn-piperidine moiety and Glu89. The predicted hydrogen bonds are in accordance with the observed structure-activity relationship in which the carboxylic acid and nitro group were shown as important pharmacophores to maintain the LIN28 inhibitory activity. The interaction involving Glu89 was the main difference predicted in binding between compound 5 and the other  spirocyclic chromenopyrazoles. To validate the docking analysis, especially regarding the interaction with Glu89, we tested compound 5 in the BLI using CSD E89A mutant, which showed significantly decreased binding affinity ( Figure S2). We followed with the cellular evaluation for the obtained spirocyclic compounds. We first determined the expression level alteration of the let-7 targeted downstream oncogene products, MYC and RAS proteins, in the human choriocarcinoma cell line JAR. As shown in Figure 4A and 4B, the endogenous levels of MYC were decreased by about 50 % with 20 μM of compound 5. RAS expression was also decreased, albeit to a less extent of about 25 % with 20 μM of compound 5. The effect of LIN28 inhibition on let-7 maturation was evaluated by qPCR (Figure 4C). Inhibition of LIN28 theoretically increases maturation of the miRNA. Compound 5 induced an increase in mature levels of two let-7 family members, let-7a and let-7 g, with up to a 3fold increase of let-7g using 20 μM of compound 5. Furthermore, to evaluate the antiproliferative effects of the spirocyclic compounds in LIN28-expressed cancer cells, we determined the effect on cell growth of compounds 2-5 in JAR cells. The results indicated that all tested spirocyclic compounds possessed single-digit micromolar inhibitory potency (IC 50 ) against the examined cell line ( Figure 4D-4E, Figure S3). Treatment with 25 μM of compound 1 did not show a significant difference in confluency after 72 h when compared to DMSO, while compound 5 at the same concentration completely inhibited cell growth and induced apoptosis of the cancer cells ( Figure 4E and Figure S3).
In addition to the role that LIN28 plays in the posttranscriptomic regulations, LIN28A has been described to bind to DNA, preferably at mismatch positions, recruiting TET1 demethylase and thereby modulating DNA methylation and gene regulation. The LIN28-DNA interaction was discovered in cells and reproduced in an EMSA experiment. [4a] Therefore, we employed the EMSA to test compound 5 measuring the interruption of the interaction between LIN28A and mismatched DNA ( Figure 5). Compound 5 inhibited complex formation in a concentration-dependent manner, similar to that of disrupting the LIN28-let-7 interaction. Based on the result, it will be interesting to study the epigenetic regulation effect of the spirocyclic chromenopyrazole LIN28 inhibitors.

Conclusions
In this work, we applied a spirocyclization strategy to develop LIN28 CSD inhibitors based on the chromenopyrazole scaffold. The resulting compounds 2-5 showed single-digit inhibitory activities against LIN28 and provided an amenable position for further appendage of other functional groups, such as a Lys-  targeting covalent warhead. The validated spirocyclic compounds exhibited potent anti-proliferation activity against JAR cells in a dose-dependent manner. Compound 5 was found to bind to LIN28A with higher affinity when compared to compound 1, induced a decrease in let-7 target gene expression, and increased the mature let-7 levels in choriocarcinoma cells expressing LIN28. In summary, "flat" molecules may not yield desirable effect to target LIN28 as most accessible binding sites are shallow and discrete. Therefore, enriching the complexity and three-dimensional nature of the small molecules via the spirocyclization strategy has the potential to facilitate the spatial extension and access more potential binding sites. This hypothesis was underlined by our docking result of compound 5 to the LIN28A cold shock domain, where additional amino acids were involved in the interaction when compared to the binding mode of let-7 RNA. Compound 5 further inhibited LIN28A binding to mismatched DNA with implications on epigenetic gene regulation. The identification of new small-molecule LIN28 inhibitors will facilitate biological understanding regarding the regulatory mechanisms involved in the LIN28-let-7 interaction. Overall, our work provides a paradigm for introducing the extendable moiety in designing RBP inhibitors via the spirocyclization strategy, which will facilitate the development of new-generation inhibitors to probe the biological functions of RBPs.
Thermal shift assay (nanoDSF). Measurements were performed with a NanoTemper Prometheus NT.48 nanoDSF instrument for assessment of melting temperatures of CSD-compound complexes. Compounds (75 μM, 5 % DMSO) were incubated with the LIN28A CSD (residues 16-126, 30 μM) for 45 min in nanoDSF buffer (30 mM NaH 2 PO 4 , pH 8.0, 50 mM NaCl, 1 mM MgCl 2 ). A temperature scan ranging from 20°C to 90°C with a slope of 1°C/min with an excitation power of 100 % was performed. The ratio of intrinsic tryptophan and tyrosine fluorescence at 350 nm and 330 nm was measured and the first derivative was determined using the device software.
Cell culture. Choriocarcinoma cell line JAR [10] was purchased from DSMZ (German Collection of Microorganisms and Cell Cultures, ACC 462) and cultured in RPMI-1640 medium supplemented with 15 % FBS and antibiotic-antimycotic solution (Sigma-Aldrich). Cells were grown at 37°C and 5 % CO 2 .
Western blot. JAR cells were treated with compounds for 24 h in a 6-well plate and cells were lysed using ice-cold RIPA buffer supplemented with SIGMAFAST™ Protease Inhibitor Cocktail (Sigma-Aldrich). Lysates were cleared by centrifugation and the protein concentration was determined using Pierce™ BCA Protein Assay (Thermo Fisher Scientific). 10-30 μg total protein were separated by SDS-PAGE followed by transfer to a nitrocellulose membrane. The membrane was blocked with 5 % dried milk in TBS-T and subsequently incubated with primary antibodies at 4°C overnight ( GAPDH, Invitrogen AM4300; Ras, Cell Signalling Technologies #3965; c-Myc, Cell Signalling Technologies #5605), washed with TBS-T and incubated with appropriate secondary HRP-coupled antibody.
MTT assay. JAR cells were seeded in 96-well plates with 3000-5000 cells/well and were incubated for a duration of 8-24 h. The medium was then discarded and the media with compounds were added in 96-well plates. Data was normalized to the medium with 1 % DMSO. The cells were then incubated for 72 h, or at 3, 6, 12, 24, 48 and 72 h in the time-dependent tests. MTT solution (5 mg/mL, 20 μL) was added in dark and incubated for additional 4 h. The old medium with MTT solution was then removed and 150 mL DMSO was added per well. The absorbance of the well was measured at 492 nm by a TECAN plate reader. Anti-proliferation assay (Incucyte). Confluency and cell death were analysed at 37°C and 5 % CO 2 using Incucyte® ZOOM (Sartorius). JAR cells were seeded into a 96-well plate (5000 cells per well) and incubated at 37°C, 5 % CO 2 . After 24 hours, cells were treated with compounds with concentrations ranging from 50 μM to 0.78 μM (0.7 % DMSO) and propidium iodide was added to the medium to monitor apoptosis. Cells were imaged for 72 hours after treatment.
qPCR. JAR-cells were grown in 6-well plates and treated with compound with a final DMSO concentration of 0.5 % for 24 h in triplicates. Total RNA was isolated using RNeasy Mini kit (Qiagen). RNA was reverse transcribed (20 ng total RNA) and analysed by qPCR using TaqMan microRNA Reverse Transcription Kit and TaqMan Universal Master Mix II with UNG (Applied Biosystems) following the manufacturers protocols. Commercial TaqMan micro-RNA assays were purchased from Applied Biosystems (Assay IDs: 001093, 000377, 002282). Signal was detected in a CFX Connect Real-Time PCR System (Bio-Rad Laboratories) and data was analyzed using the 2 À ΔΔCT method normalized to U6 snRNA.

Docking analysis.
For computational docking analysis of spirocyclic chromenopyrazoles to the preE-let-7 binding site of LIN28A CSD (PDB code: 5UDZ). [11] Schrödinger® Maestro 12.3 was used. The three-dimensional structures of compounds were prepared after calculating energy minimization by MM2 with PerkinElmer Chem3D® 20.1. Chemical states were generated with the ligand preparation module and LIN28A conformation with the protein preparation module. Latter included hydrogen addition, water molecule removal and energy minimization. To reveal crucial interactions of preE-let-7 with LIN28A in order to identify binding sites for the spirocyclic compounds, fragments of the miRNA were removed except for oligomers used as ligands. The binding site for docking was then generated by the grid generation module and using preE-let-7 fragments (A4-U8, U8-U11 or U13-A15) or crucial residues of LIN28A involved in the binding of preE-let-7. The glide dock module was used and the results were evaluated according to interactions between small molecules and LIN28A, small molecule orientations, docking scores and solvent exposure patterns. The interactions of spirocyclic chromenopyrazoles and LIN28A were visualized using PyMOL 2.4.1.

Supporting Information
Additional references cited within the Supporting Information. [12]