Identification of a Nonphosphorylated, Cell Permeable, Small Molecule Ligand for the Stat3 SH2 Domain

Abstract

Signal transducer and activator of transcription 3 (Stat3) protein is a cytosolic transcription factor that is aberrantly activated in numerous human cancers. Inhibitors of activated Stat3–Stat3 protein complexes have been shown to hold therapeutic promise for the treatment of human cancers harboring activated Stat3. Herein, we report the design and synthesis of a focused library of salicylic acid containing Stat3 SH2 domain binders. The most potent inhibitor, 17o, effectively disrupted Stat3:phosphopeptide complexes (K_i = 13 μM), inhibited Stat3–Stat3 protein interactions (IC₅₀ = 19 μM) and silenced intracellular Stat3 phosphorylation and Stat3-target gene expression profiles. Inhibition of Stat3 function in both breast and multiple myeloma (MM) tumor cells correlated with induced cell death (EC₅₀ = 10 μM and 16 μM, respectively).

1. Introduction

As a master regulator of oncogenic cellular processes, signal transducer and activator of transcription (Stat) 3 protein has become the focus of molecularly targeted anti-cancer therapeutic development. Stat3 protein is a cytosolic transcription factor that plays a key role in mediating cell division and apoptosis.^1,2 Stat3 signaling is initiated by extracellular cytokine³ or growth factor⁴ receptor stimulation and results in the expression of anti-apoptotic proteins that control cell growth and survival.^2,5 As part of the signaling cascade, Stat3 is recruited to the intracellular domain of the target receptor, where it is phosphorylated on Tyr705.^6–8 Once phosphorylated, Stat3 dissociates, binds another activated Stat3 monomer through reciprocal Src Homology 2 (SH2) domain-phosphotyrosine interactions and translocates to the nucleus. Once in the nucleus, dimeric Stat3 binds to DNA and promotes the transcription of proteins that govern cell cycling and prevent apoptosis.^5–8 In healthy cells, Stat3 activity is transient and tightly controlled by supressors of cytokine signaling, phosphpatases and proteosomal degradation.⁹ In many human cancers, however, Stat3 activity is hyperactivated leading to overexpression and accumulation of anti-apoptotic proteins within the cell. Elevated levels of Stat3 activation confer resistance to natural apoptotic cues and allows for rapid proliferation and de novo tumorogenesis. Aberrantly activated Stat3 is found in numerous human cancers including leukemia and lymphoma, as well as cancers of the breast, prostate, lung, head, neck, and ovaries.^2,9–11

Numerous studies have demonstrated that inhibition of Stat3 activation leads to reduced levels of Stat3-target gene expression profiles and correlates with programmed cell death.^12–14 To date, effective disruption of Stat3 function has been achieved primarily through inhibition of transcriptionally active Stat3–Stat3 dimers. The Stat3–Stat3 binding complex is characterized by large, non-contiguous intrafacial surface areas possessing few targetable binding sites.¹² As a result, the development of potent small-molecule Stat3 inhibitors remains a challenging task. The majority of published Stat3 inhibitors bind Stat3’s phosphopeptide binding SH2 domain (Figure 1).^{11–13,15–17}

Figure 1.

Small molecule Stat3 SH2 domain binders.

We have recently identified a potent salicylic acid-based Stat3 inhibitor, 7 (SF-1-066⁶ (or 27h)¹⁸) after a structure activity relationship (SAR) study of compound 1 (S3I-201, Figure 1). Inhibitor 7 showed promising anti-Stat3 activity both in vitro, disrupting Stat3 protein–phoshopeptide and Stat3–Stat3 protein–protein interactions and elicited in vivo suppression of breast tumor xenografts.¹⁹ Moreover, fluorescence polarization binding experiments showed that 7 is selective for Stat3’s SH2 domain cf. Stat5 and Stat1 isoforms (Stat3 K_i = 15 μM; Stat5, K_i > 25 μM; Stat1, K_i > 25μM). Encouragingly, 7 showed negligible effects against ‘healthy’ cells lacking activated Stat3 (NIH3T3, TE-71 and HPDEC) and selectively killed cancer cells harboring aberrant Stat3 activity.¹⁹ GOLD²⁰ docking studies revealed that compound 7 binds to the pTyr-binding portion of the SH2 domain, with the salicylic acid making interactions with Lys591, Glu594 and Arg609.^6,19 In addition, the hydrophobic cyclohexylbenzyl appendage forms van der Waal’s interactions with a series of predominantly hydrophobic residues (Figure 2).

Figure 2.

GOLD docking studies of 7 bound to Stat3’s SH2 domain.⁶

In this study we investigated the binding significance of the toluene-sulfonamide substituent to Stat3-SH2 domain recognition. We herein report an SAR of the sulfonamide portion of compound 7, and present novel analogs, including 17o, which exhibited improved inhibition of Stat3 function both in vitro and in whole cell tumor models of breast and multiple myeloma cancers.

2. Materials and Methods

2.1 Electrophoretic Mobility Shift Assay

EMSA analysis was performed as previously reported.^6,19 Nuclear extracts of NIH3T3/vSrc cells were pre-incubated with varying concentrations of compounds for 30 min at room temperature prior to incubation with ³²P-labeled oligonucleotide probe, hSIE (high affinity sis-inducible element from the c-fos gene, m67 variant, 5′-AGCTTCATTTCCCGTAAATCCCTA) for 30 minutes at 30 °C before subjecting to EMSA analysis. DNA-binding activities were measured for each band at each concentration of inhibitor and quantified using ImageQuant. Results were plotted as percent of control from which an IC₅₀ value could be derived.

2.2 Fluorescence Polarization Assay

As previously reported,^6,21 fluorescence polarization experiments were performed on an Infinite M1000 (Tecan, Crailsheim, Germany) using black 384-round bottom well plates (Corning), and buffer containing 50 mM NaCl, 10 mM Hepes, pH 7.5, 1 mM EDTA, and 2 mM dithiothreitol and a final concentration of 5 % DMSO. Stat3 protein (150 nM) was treated with varying concentrations of inhibitor compounds (100 to 0.19 μM final concentrations). The fluorescent probe was added at a final concentration of 10 nM. Protein, inhibitor and probe were combined and incubated for 15 minutes prior to analysis. Polarized fluorescence was plotted against concentration and fitted using a standard dose response curve. K_i values were calculated using the formula below where [STAT3] = 150 nM and K_d = 150 nM.

$$K_i=\frac{{IC}_{50}}{1+\frac{[\mathit{STAT}3]}{K_d}}$$

2.3 Whole Cell Cytotoxicity Assays

Human cell lines, DU145, OCI-AML2 and JJN3 were prepared in 96 well plates and treated with varying concentration of inhibitor. After 72 hours, cell growth and viability was measured with the CellTiter96 aqueous nonradioactive (MTS) assay according to the manufacturer’s instructions (Promega, Madison, WI) and as described previously.²² Relative viability was plotted versus concentration and EC₅₀ was determined by fitting to a standard dose response curve.

2.4 Immunoblotting

Cells were lysed in lysis buffer (50 mM Tris-HCl, 1 mM EDTA, 1% NP-40, 150 mM NaCl) for 30 minutes on ice, then at −80 °C freeze/thaw once and clarified by centrifugation at 12000g for 15 minutes. Proteins were separated by 6.5% to 15% sodium dodecyl–polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotted with the specified antibody. Protein bands were visualized using secondary antibodies coupled to horseradish peroxidase and the Chemiluminescence Reagent Plus (from Perkin Elmer Life Sciences) according to the manufacturer’s instructions. Anti-cMyc was purchased from Santa Cruz, anti-survivin from NOVUS Biologicals, Anti-Mcl-1, and anti-Bcl-xL from BD Biosciences, (Mississauga, ON), anti-phospho STAT3, anti-STAT3 and anti-PARP are from Cell Signaling Technology, (Pickering, ON).

3. Results and Discussion

A family of 16 novel sulfonamide analogs of 7 were prepared as outlined in Scheme 1. Briefly, 4-aminosalicylic acid (8) was doubly benzylated in one pot using potassium tert-butoxide and benzyl bromide. Next, aniline 9 was reductively aminated with 4-cyclohexylbenzaldehyde using NaCNBH₃. In parallel, we TFA protected the amino group of sarcosine tert-butyl ester (11) to furnish 12, and then removed the tert-butyl ester under acidic conditions (TFA/CH₂Cl₂) to yield the carboxylic acid, 13. Condensation of 13 with secondary aniline 10 furnished tertiary amide, 14. The TFA protecting group was then removed by LiOH mediated hydrolysis revealing secondary amine, 15. In the penultimate step we coupled a diverse variety of sulfonyl chlorides to 15, yielding compounds, 16a-o. Finally, hydrogenolysis conditions (H₂, 10% Pd/C) were employed to debenzylate the salicylic acid moiety, exposing final compounds 17a-o. Of note, in cases where hydrogenolysis conditions were incompatable with the sulfonyl substituent (17f, 17j, 17k and 17n), we employed a step-wise, TFA mediated debenzylation of phenol, followed by LiOH hydrolysis of the benzyl ester (Scheme 1, steps i, j).²³

Scheme 1.

a) BnBr (2eq), KO^tBu, DMF, 0 °C, 16 h, 73%; b) 4-cyclohexylbenzaldehyde, AcOH, NaCNBH₃, rt, 16 h, 79%; c) (CF₃CO)₂O, DIPEA, CH₂Cl₂, rt, 3 h, 96%; d) TFA/CH₂Cl₂, 1:1, rt, 5 h, 100 %; e) 10, PPh₃Cl₂, CHCl₃, 60 °C, 12 h, 97%; f) LiOH·H₂O, THF/H₂O, 3:1, rt, 10 min, 98%; g) RSO₂Cl, DIPEA, CH₂Cl₂, rt, 16 h, 78–98% h) H₂, 10% Pd/C, MeOH/THF, 1:1, rt, 1–16 h, 85–100%; or for 17f, 17j, 17k and 17n: i) LiOH·H₂O, THF/H₂O, 3:1, rt, 24 h, 73–89%; j) TFA/CH₂Cl₂, 1:2, rt, 16 h, 65–92%.

We first assessed for inhibitor induced Stat3–Stat3 dimer disruption using a routinely used Electrophoretic Mobility Shift Assay (EMSA), which measures dimer disruption through inhibition of DNA binding.²⁴ As illustrated in Table 1, varying the sulfonamide substituent resulted in varying degrees of inhibition potency. We incorporated a range of appendages to cater for the relatively amphiphilic pocket composed of residues Ile634, Ser636, Glu594 and the hydrophobic chain of Lys591. In general, hydrophobic R groups afforded the most potent inhibitors. The polar 17h, incorporating a 1-methyl-1H-imidazole group, lost all inhibitory potency (IC₅₀ > 300 μM). Interestingly, employing the meta-tolyl isomer 17a significantly reduced activity, (17a (meta-) IC₅₀ = 118.8 μM cf. 7 (para-) IC₅₀ = 35 μM)). The bulkier 2,4,6-tri-methylphenyl substituted inhibitor, 17b exhibited weaker activity than the parent compound 7, with an IC₅₀ = 51.9 μM. The larger biphenyl sulfonamide, 17c, was a modest inhibitor of Stat3 dimerization (IC₅₀ = 65.4 μM), as was the 2-naphthyl derivative, 17d (IC₅₀ = 79.2 μM). Notably, bis-aryl sulfonyl derivatives substituted at the 1-position, including, 17e (R = 1-naphthyl, IC₅₀ = 28.8 μM), 17f (R = 8-quinolinyl, IC₅₀ = 25 μM) and 17g (R = dansyl, IC₅₀ = 29 μM) proved to be active Stat3 inhibitors. Taken together, these data suggest that substitution of the ortho- and meta-tolyl positions of 7, with a second aryl group is better tolerated than in the para position. In general, replacement of the methyl group in the para-tolyl moiety of 7 with different isosteres (F, Br, Cl, OMe, NO₂) led to a reduction in Stat3 inhibitory activity. However, 17o, incorporating a pentafluorophenyl sulfonamide substituent, proved to be the most active of the phenyl sulfonamide series. Indeed, 17o was approximately two-fold more potent as the parent compound 7 (17o, IC₅₀ = 19 μM cf. 7, IC₅₀ = 35 μM).

Table 1.

EMSA inhibition data for the disruption of the Stat3–Stat3:DNA complex by sulfonamide analogs 7 and 17a–17o.

inhibitor	R	IC50 (μM)
7		35 ± 9
17a		118.8 ± 1.9
17b		51.9 ± 2.4
17c		65.4 ± 7.1
17d		79.2 ± 11.2
17e		28.8 ± 2
17f		24.6 ± 3.4
17g		28.8 ± 1.9
17h		>300
17i		126.2 ± 5.3
17j		90.3 ± 4.5
17k		67.2 ± 2.6
17l		>300
17m		67.4 ± 4.9
17n		62.2 ± 3.2
17o		19.7 ± 5.4

Next, we investigated the binding potency of select agents against Stat3’s SH2 domain via a routinely used fluorescence polarization assay, the results of which are shown in Table 2. Encouragingly, compounds 17b (K_i = 8.0 μM), 17c (K_i = 6.2 μM), 17g (K_i = 13.3 μM), 17k (K_i = 11.0 μM) and 17o (K_i = 12.8 μM) exhibited improved activity compared to compound 7. Although similar trends were observed in the EMSA and FP data, there are some notable deviations between the two data sets of data. For example, compound 17b, IC₅₀ = 51.9 μM in the EMSA assay is significantly more potent in the FP assay (K_i = 8.0 μM). As previously reported,⁶ this anomaly between EMSA analysis of nuclear extracts, and the FP assay is likely due to the presence of other Stat isoforms and proteins found in the nuclear extracts. Taken together, the EMSA and FP results suggest that we are able to disrupt Stat3:phoshopeptide and Stat3–Stat3 complexation events by effectively blocking the Stat3 SH2 domain.

Table 2.

Fluorescence polarization assay binding data (K_i values in μM)

inhibitor	R	Ki (μM)
7		15.5 ± 4.7
17b		8.0 ± 2.4
17c		6.2 ± 2.0
17e		26.5 ± 0.4
17f		41.0 ± 0.4
17g		13.3 ± 0.6
17h		> 100
17i		30.6 ± 11
17k		11.0 ± 0.3
17o		12.8 ± 0.3

Since blockage of Stat3 signaling in compromised cell lines leads to induced apoptosis,²⁵ we reasoned that our most potent inhibitors would kill cells harboring activated Stat3. Thus, we employed an MTS assay to assess the whole cell potency of select inhibitors including, 7, 17e and 17o, which showed activity in both EMSA and FP-assays.^26,27 DU145 (prostate), MDA-468 (breast) and JJN3 (multiple myeloma) cancer cells were incubated for 72 hrs with varying concentrations of inhibitors and relative viability assessed colorometrically after treatment with MTS for 3 hrs. Notably, compound 17o displayed an approximately two-fold increase in potency over 7, with IC₅₀ values of 10, 23 and 17 μM in breast, prostate and MM cells, respectively. Compound 17e showed lower activity in cells than both 7 and 17o, possibly a result of increased lipophilicity and poorer water solubility (17e, logP = 5.95 versus 17o, logP = 5.74). We noted that 17o exhibited much improved water solubility over both 7 and 17e. While the in vitro activities of 17o and 7 are comparable, we postulated that the resultant increase in cellular activity may be a result of greater cell permeability and reduced aggregation/precipitation.

Due to the promising cytotoxic activity observed in tumor cells, 17o was assayed for inhibition of Stat3 phosphorylation in both MDA-468 and JJN3 cell lines harboring activated Stat3.^28,29 As a control, Western blot analysis showed that control inhibitor, 7 effectively knocked down Stat3 phosphorylation at approximately 100 μM in both MDA-468 and JJN3 cancer cells. Most encouragingly, 17o inhibited Stat3 phosphorylation at significantly lower concentrations (20 μM) in intact cells after 24 hrs. Furthermore, immunoblotting analysis of the same cell lines after the same time period revealed that 17o effectively reduced levels of Stat3 downstream targets, including, cMYC, Bcl-xL and Survivin. We presume that the resultant cytotoxicity observed after 72 hr incubation is a result of 17o/7-induced inhibition of intracellular Stat3 signaling. The data shows that 17o is a more potent whole cell inhibitor of Stat3 function than lead compound, 7, presumably due to improved solubility and cell permeability. We will conduct further investigations to elucidate the biological and biochemical mechanisms of 17o’s improved anti-cancer activity which will be published elsewhere.

4. Conclusion

We have presented the design and synthesis of a novel family of Stat3 inhibitors that exhibit promising in vitro binding potency for the Stat3 SH2 domain, as well as improved tumor whole cell activity. Most notably, hit compound, 17o, showed an approximately 2–4-fold increase in in vitro activity compared to lead agent, 7, and nearly 6-fold higher potency in JJN3 MM tumor cells. Future studies seek to evaluate the in vivo properties of 17o in MM and breast tumor xenograft models. Thus, to date, 17o represents the most potent Stat3 inhibitor derived from the salicylic acid-based class of inhibitors.

Figure 3.

SDS-Page and Western blotting analysis of whole cell lysates prepared from MDA-468 human breast cancer and multiple myeloma JJN3 cells, untreated (DMSO, control) or treated with 17o (15 μM), 17 e (125 or 150 μM), and 7 (100 or 125 μM) for 24 h and subjected to immunoblotting analysis for pY705Stat3, Stat3, c-Myc, Bcl-xL, Mcl-1 and Survivin.

Table 3.

Whole cell MTS data. Cells were treated with varying concentrations of inhibitors for 72 hours.

inhibitor	R	EC50 (μM)
		MDA-468	DU145	JJN3
7		17.0 ± 4.4	37.2 ± 12.4	93.3 ± 15.8
17e		46.5 ± 12.4	74.5 ± 30.2	106.1 ± 13.7
17o		10.9 ± 3.0	22.7 ± 8.5	16.7 ± 0.7