Synthesis and Evaluation of Novel 1,2,6-Thiadiazinone Kinase Inhibitors as Potent Inhibitors of Solid Tumors

A focused series of substituted 4H-1,2,6-thiadiazin-4-ones was designed and synthesized to probe the anti-cancer properties of this scaffold. Insights from previous kinase inhibitor programs were used to carefully select several different substitution patterns. Compounds were tested on bladder, prostate, pancreatic, breast, chordoma, and lung cancer cell lines with an additional skin fibroblast cell line as a toxicity control. This resulted in the identification of several low single digit micro molar compounds with promising therapeutic windows, particularly for bladder and prostate cancer. A number of key structural features of the 4H-1,2,6-thiadiazin-4-one scaffold are discussed that show promising scope for future improvement.

It is important to note that the stability of dianilinothiadiazinones in biological systems has previously been assessed [15]. 3,5-Bis(phenylamino)-4H-1,2,6-thiadiazin-4-one [18] is stable to neutral, acidic, or slightly basic aqueous conditions, the presence of amine or thiol nucleophiles, and oxidizing and reducing conditions. We similarly tested thiadiazinone 16 from our current study and found the same results.

Cancer Cell Screening
Compounds 14-26 were screened on an array of cancer cell lines to explore the structural requirements for anti-cancer activity within the 1,2,6-thiadiazinone scaffold (Table 1). This included pancreatic, bladder, prostate, breast, chordoma, and lung cancer cell lines while a skin fibroblast cell line was used as a toxicity control [33][34][35][36]. These cancer cell lines present different drug resistance profiles and when combined in a panel, each represents a distinct therapeutic challenge to overcome [37].
The first analogue 3-[(3-acetylphenyl)amino]-5-(1H-pyrrolo[2,3-b]pyridin-4-yl)-4H-1,2,6-thiadiazin-4-one (14) demonstrated limited inhibition across the panel of cell lines with the most inhibition shown on the bladder cancer cell line (IC 50 = 15 µM). Switching from the acetyl substitution 14 to the methoxy 15 showed no improvement across the panel. Removal of the methyl on the methoxy 15, to afford a hydroxy functionality 16, led to a 10-fold increase in potency on the bladder cancer cell line (IC 50 = 1.6 µM) with no toxicity in the WS-1 cell line (IC 50 = >100 µM). The introduction of a 4-methyl substituent 17 removed weak inhibition on PANC1 and UCH-1 but did not show any improvement on the other cell lines in the panel. Interestingly, the introduction of a 2-methyl substituent 18 removed nearly all anti-cancer activity. Switching to the aliphatic a morpholine substituent again changed the preference toward pancreatic (IC 50 = 11 µM) and prostate (IC 50 = 14 µM) cancers with little activity towards bladder and the other cell lines.
We then sought to look at changing the hinge binding motif using the methyl piperazine as the fixed substituent. This tertiary amine substitution affords compounds with favorable properties [38][39][40][41][42]. However, in the case of our three powerful hinge binders [30], we observed very limited activity across both azaindoles, 4-postion 20 and 6-position 21, along with the indazole 22. The indazole 22 showed a small amount of toxicity in the WS-1 cell line (IC 50 = 24 µM).

Kinome Profiling of Thiadiazinones 16, 17, and 26
We then assessed the kinome profile of thiadiazinones 16, 17, and 26 all at 1 µM using a multiplexed kinase inhibitor bead set and quantitative mass spectrometry for detection of kinase peptides (MIB-MS) [36,45,46]. The MIB-MS proteomics was able to detect between 350 and 400 kinases in the cell lysates and competitive displacement of specific kinases from the beads was used to determine the kinome profile of the thiadiazinone kinase inhibitors. Compound 16 showed a narrow spectrum kinome profile with weak interactions on MAP4K5, MAP4K3, PRKCD, and PKN1 just below the threshold ( Figure 3). Compound 17 also had a narrow spectrum kinome profile, with AKT2, MAP2K4, MAP2K1, and STK35 being just below the threshold ( Figure 4). Compound 26 had a narrow spectrum kinome profile but was a weak inhibitor, showing affinity for ACVR1B, ACVR2A, and STK35 ( Figure 5).  The key hits on thiadiazinones 16, 17, and 26 from the MIBS profiling, were then modelled to understand their interactions in the respective ATP binding site (Figures 6-8). First, compound 16 was analyzed against MAP4K5 [47], MAP4K3 [48], PRKCD [49], and PKN1 [50]. Docking of the compounds was performed using Schrödinger Maestro [51]. Before docking into the ATP-binding site, prepared from the PDB structure as needed, the compounds were minimized using LigPrep. The docking pose that placed the azo-indole in a position to make a hydrogen bond from the backbone NH of the hinge region of the respective kinases was the most favorable ( Figure 6A-D). The 1,2,6-thiadiazinone is relatively passive, acting as a linker to the solvent region where the phenolic alcohol can form a series of different hydrogen bonding interactions. In the case of MAP4K5, MAP4K3, and PRKCD ( Figure 6A-C), this is both accepting and donating, forming a rigid fixture at the mouth of the ATP binding site. PKN1 is not able to accommodate this accepting interaction, so there is only a much weaker hydrogen bonding interaction with D764 and no interaction with K748. In the case of compound 17 (Figure 7), the azo-indole was similarly positioned at the hinge in both AKT2 [52] and STK35 [47,53]. However, the rest of the molecule was not orientated differently in both cases. This enabled additional interactions in both cases where the 1,2,6-thiadiazinone is actively involved, forming a hydrogen bond with the residue T292 in AKT2 and D323 in STK35. This is in addition to a hydrogen bond interaction by the phenolic alcohol with N280 and N365 in AKT2 and STK35, respectively.
Thiadiazinone 26 (Figure 8), similarly to compounds 16 and 17, formed a strong hydrogen bond in the hinge region of ACVR1B [54] and STK35 [53]. In AVCR1B, the 1,2,6-thiadiazinone was passive with no strong interactions. The pyridine nitrogen did, however, form a hydrogen bond with a solvent-exposed lysine residue K337. In the case of STK35, the 1,2,6-thiadiazinone was actively involved, forming an interaction with a solvent-exposed lysine residue K231. This was in addition to another hydrogen bond interaction with the bringing amine between the indazole and 1,2,6-thiadiazinone. The ATP was too shallow to accommodate the pyridine interaction as seen in AVCR1B and interactions involved with the phenolic alcohol of compound 17.

Discussion
Previously, we demonstrated that the 4H-1,2,6-thiadiazin-4-one chemotype can function as an ATP-competitive kinase inhibitor, acting as a novel hinge binding motif. This was also the first report of a protein co-crystallization with this rare heterocycle [15]. We have now expanded the repertoire of the 4H-1,2,6-thiadiazin-4-one as a spacer unit within a kinase inhibitor affording unique chemical properties. The electronics of the 4H-1,2,6thiadiazin-4-one core allows for negative charge transfer from the sulfur atom towards the conjugated system [55]. This electronic property, exploited in solar cell applications, can partly explain the general lack of kinome promiscuity compared to the dianilinopyrimidine [15]. The modular synthesis and relative narrow kinome spectrum make the 4H-1,2,6-thiadiazin-4-one an attractive chemotype for further development.
The chemical tractability of the protein kinases makes them an attractive target for drug development. Over 90 drugs have been approved that target the ATP binding site, predominantly in the field of oncology [19,56]. However, with emerging indications beyond cancer including the treatment of chronic diseases, such as inflammation and neurodegeneration, along with the emergence of resistance, development of compounds with improved potency and selectivity profiles are essential to remain at the cutting edge [57,58]. A number of indirect assays exist to assess the inhibitor affinity, and binding assays are some of the most accurate and robust methods to measure the potency and selectivity of ATP-competitive kinase inhibitors [59][60][61][62][63]. Ligand binding displacement assays are particularly important to provide an accepted form of direct measurement of kinase inhibition in drug optimization of ATP-binding site inhibitors for neglected kinases, where there are currently no robust and validated enzyme activity assays [59,63].
Our results afford new data on a series of interesting starting points for further optimization towards new kinases targets including MAP4K3, MAP4K5, PRKCD, PKN1, AKT2, STK35, and ACVR1B, all of which have some indication in proliferation. The early kinome data, combined with modelling and the accompanying phenotype on bladder cancer, afford exciting prospects for this scaffold. This work firmly puts 1,2,6-thiadiazin-4-one into the medicinal chemistry toolbox and provides another example of the application of sulfur heterocycles in drug design.

MIBS Profiling Methods
Previously described [45]. Briefly, multiplexed inhibitor beads (MIBS) affinity chromatography/MS analysis: SUM159 cells were cultured in a 50:50 mixture of DMEM and Nutrient Mixture F-12 medium (Gibco) supplemented with 5% fetal bovine serum, 1% anti/anti, 5 mg mL −1 insulin, and 1 mg·mL −1 hydrocortisone. Cells were maintained at 37 • C in a humidified 5% CO 2 atmosphere. SUM159 cells were grown to 80% confluency, washed twice with PBS, and harvested by scraping cells in lysis buffer [50 mM HEPES, pH 7.5, 150 mM NaCl, 0.5% Triton X-100, 1 mM EDTA, 1 mM EGTA, 10 mM NaF, 2.5 mM Na 3 VO 4 , complete protease inhibitor cocktail (Roche, Basel, Switzerland), phosphatase inhibitor cocktail 2, and 3 (Sigma, St. Louis, MO, USA)]. Lysates were sonicated then clarified by centrifugation at 14000 V g for 15 min at 4 • C. Lysate was then filtered through a 0.2 mm syringe filter and frozen at −80 • C until use. Protein concentration was quantified using a Bradford assay the day of the experiment. DMSO or the indicated concentration of 16, 17, and 26 were added to lysate containing 4 mg of total protein. Lysates were vortexed briefly then incubated for 30 min on ice. Kinases were affinity purified as previously described [46]. Briefly, lysates were diluted to 1.33 mg mL −1 with lysis buffer then NaCl concentration brought to 1 M. Diluted lysates were passed over a mixture of 25 mL of settled beads of each of the following inhibitors conjugated to ECH Sepharose beads: Purvalanol B, PP58, VI-16832, UNC21474A, UNC8088A, and 37.5 mL of settled beads conjugated to CTx-0294885 and VI-16832 [69]. The kinase inhibitor-bead conjugates were previously equilibrated in high-salt buffer (50 mM HEPES pH 7.5, 1M NaCl, 0.5% Triton X-100, 1 mM EDTA, and 1 mM EGTA). MIBs columns were sequentially washed with high-salt buffer, lowsalt buffer (50 mm HEPES pH 7.5, 150 mM NaCl, 0.5% Triton X-100, 1 mM EDTA, and 1 mM EGTA), and SDS buffer (50 mm HEPES pH 7.5, 150 mm NaCl, 0.5% Triton X-100, 1 mm EDTA, 1 mm EGTA, and 0.1% SDS). Proteins were eluted by boiling samples in elution buffer (100 mM Tris·HCl pH 6.8, 0.5% SDS, and 1% β-mercaptoethanol) for 15 min twice. Dithiothreitol (DTT) was added to a final concentration of 5 mm and samples were incubated at 60 • C for 25 min. Samples were then cooled to room temperature on ice and alkylated by adding iodoacetamide to a final concentration of 20 mm for 30 min in the dark at room temperature. Samples were then concentrated in 10K Amicon Ultra centrifugal concentrators (Millipore, Burlington, MA, USA) followed by methanol and chloroform precipitation of proteins. The final protein pellets were re-suspended in 50 mM HEPES pH 8.0 and incubated with trypsin at 37 • C overnight. Residual detergent was removed by three sequential ethyl acetate extractions then desalted using Pierce C 18 spin columns (Thermo Scientific, Waltham, MA, USA) according to the manufacturer's protocol.

Ligand Preparation
Structures of small molecules were parametrized and minimized using the LigPrep module of Schrodinger 2020-4 suite employing OPLS3e force field [51].

Development of Homology-Based Model
Three-dimensional models of ACVR1B, STK35, and MAP4K5 were developed using Biovia Discovery Studio 2019 program (Dassault Systèmes, San Diego, CA, USA). The human forms of target sequences were downloaded from uniport, and psi-blast searched against the PDB-template database. Pre-aligned template structures (ACVR1B-(PDB:5E8X) [73], STK35-(PDB:3MA6), and MAP4K5-(PDB:5J5T) [70]) were downloaded to Discovery Studio and Alignments to corresponding kinase domains. The sequence of corresponding unknown kinase was then aligned to template structures. The models were built using standard settings of the modeler: optimization level = high, Number of models = 20, Refine loops = false, original ligands were copied from templates. Models were then ranked according to the PDF score and highest ranked models were uploaded to Maestro for protein preparation protocol. Models of AVCR1B and MAP4K5 were also successfully generated using protein structure prediction functionality available at GalaxyWEB-modelling site (http://galaxy.seoklab.org/ Accessed on: 6 September 2021), and cross checked with in-house generated models. During the final stage of this study, several AlphaFold structures available including ACVR1B, STK35, and MAPK5 were downloaded [74], prepared, and aligned to compare the positions of key residues with previously generated in-house models, which were consistent with our findings (data not shown).

Molecular Docking
The ligand docking was performed using an induced fit protocol using the standard settings of Schrodinger suite 2020-4 (up to 20 models were generated, dockings were done using SP-level of IFD-setting during Glide docking and redocking steps). The centers of glide grids were set to the centroid of co-crystallized ligands or template-derived ligands. The highest ranked docking poses were visually examined to conclude favorable hinge binding forms.