Structure–Activity Study of Nitazoxanide Derivatives as Novel STAT3 Pathway Inhibitors

Abstract

We identified nitazoxanide (NTZ) as a moderate STAT3 pathway inhibitor through immunoblot analysis and a cell-based IL-6/JAK/STAT3 pathway activation assay. A series of thiazolide derivatives were designed and synthesized to further validate the thiazolide scaffold as STAT3 inhibitors. Eight out of 25 derivatives displayed potencies greater than that of NTZ, and their STAT3 pathway inhibitory activities were found to be significantly correlated with their antiproliferative activities in HeLa cells. Derivatives 15 and 24 were observed to be more potent than the positive control WP1066, which is under phase I clinical trials. Compared with NTZ, 15 also exhibited much improved in vivo pharmacokinetic parameters in rats and efficacies against proliferations in multiple cancer cell lines, indicating a broad-spectrum effect of these thiazolides as antitumor agents targeted on STAT3.

Nitazoxanide (NTZ), or 2-(acetyloxy)-N-(5-nitrothiazol-2-yl)benzamide (Figure 1A), is an oral antiparasitic drug approved by the U.S. Food and Drug Administration (FDA) for the treatment of diarrhea caused by Cryptosporidium parvum and Giardia intestinalis infection in both adults and children.¹ It was initially synthesized in the early 1970s on the scaffold of niclosamide (NCS) (Figure 1A), which is also an antiparasitic drug used in humans to treat tapeworm infection,² by replacing one benzene ring with a five-membered thiazole ring for improved solubility and better safety profiles.³ In the recent years, both of these anthelmintic drugs have demonstrated a broad spectrum of antibacterial, antivirus, and anticancer effects in many independent studies⁴⁻⁷ and are frequently hit in parallel from high-throughput screenings aiming to repurpose old drugs for novel clinical uses,⁸⁻¹¹ indicating a close relationship between the underlying mechanisms involved in the similar bioeffects of these two drugs. Indeed, both of these compounds have been reported to be able to regulate multiple signaling pathways and biological processes,⁴ such as the Wnt/β-catenin pathway,¹² c-Myc,¹³ and mTORC1.¹⁴

Figure 1.

(A) Chemical structures of NCS, NTZ, and WP1066. (B) Relative activity of the IL-6/JAK/STAT3 pathway in HEK-Blue IL-6 cells treated with NTZ or WP1066. (C, D) Cell viabilities of HEK-Blue IL-6 cells after treatment with the indicated concentrations of (C) NTZ or (D) WP1066 for 20 h. Data are representative of three independent experiments.

As early as in 2010, Ding and colleagues¹⁵ identified NCS as an inhibitor of signal transducer and activator of transcription 3 (STAT3) that significantly inhibits activation, nuclear translocation, and transactivation of the STAT3 pathway in cancer cells. Because of the structural similarity and mounting evidence of parallel biofunctions of the two compounds, we speculated that NTZ could also have potency to regulate the STAT3 pathway and hence contributed to the understanding of the mechanisms underlying its observed anticancer effects.^12,13,16

STAT3 is a prominent member of the STATs family,¹⁷ which play dual functions of signal transduction and transcription activation by transmitting extracellular signals into the nucleus and modulate the expression of specific target genes¹⁸ involved in the regulation of many fundamental cellular processes, including growth, survival, proliferation, differentiation, and immune responses.^19,20 Under normal physiological conditions, the STAT3 signaling is strictly controlled at a low level by many intracellular suppressors and various protein tyrosine phosphatases.²¹ However, persistent activation of STAT3 occurs frequently in a variety of human cancers, stimulating tumor angiogenesis, promoting immune evasion, and even conferring resistance to apoptosis induced by conventional therapies.²² Therefore, inhibition of the constitutively active STAT3 pathway has been an attractive strategy for cancer therapy both by inhibition of tumor cell growth and through stimulation of antitumor immunity.²²⁻²⁴ However, identification of promising small-molecule inhibitors of this transcription factor is challenging. To date, only several STAT3 inhibitors have advanced into early-phase clinical trials, and no STAT3-targeted therapies have yet been approved to treat cancer.²⁵ In the present study, we demonstrate that NTZ is a novel inhibitor against the STAT3 pathway. Preliminary deviations on the phenyl ring were performed to further confirm that NTZ-related 5-nitrothiazolides are effective STAT3 pathway inhibitors.

To identify whether NTZ affects STAT3 activation, we first examined its inhibitory activity on a cell-based model²⁶ where upon IL-6 stimulation STAT3 is steadily activated in HEK-Blue IL-6 cells in which secreted embryonic alkaline phosphatase (SEAP) is expressed and secreted in response (Figure S1). Using a SEAP detection reagent, the activity of the STAT3 pathway can be quantitatively evaluated. Meanwhile, to avoid false-positive results introduced by a decrease in the number of viable cells, the cell viability was also measured through 3-(4,5-dimethylthiazol-2-yl)-2,3-diphenyltetrasodium bromide (MTT) assay. Compound WP1066 (Figure 1A), which is currently in phase I clinical trials to treat recurrent malignant glioma or progressive metastatic melanoma in the brain,^27,28 was adopted as a positive control because of its excellent potency against the STAT3 signaling pathway. As shown in Figure 1B, NTZ demonstrated moderate but complete inhibition of IL-6-induced STAT3 pathway activation. Although it was less potent than WP1066, its cytotoxicity was much lower as well, as no significant cell death was observed even at 80 μM (Figure 1C). In contrast, obvious cytotoxicity was observed for WP1066 when the concentration was higher than 10 μM (Figure 1D).

Encouraged by the above results, we further investigated the detailed effects of NTZ on intracellular STAT3 activation using HeLa epithelial carcinoma cells because of their constitutive overexpression of STAT3.²⁹ After treatment with NTZ for 1 h, Tyr705 phosphorylation of STAT3 was suppressed in a dose-dependent manner, whereas the total STAT3 remained unchanged (Figure 2A). Interestingly, the time-course study revealed that NTZ could also affect the phosphorylation of Ser727. As shown in Figure 2B, there was a sharp decline in the p-STAT3 (Tyr705) level from 15 min to 2 h after treatment with 80 μM NTZ, and after a 6 h interval, p-STAT3 (Tyr705) was observed to rebound temporarily, followed by a gradual decrease afterward. In contrast, a higher phospho-Ser727 level was observed within the first 30 min and then gradually declined to normal or subnormal level. This upside-down U-shaped change implied an indirect effect of NTZ on p-STAT3 (Ser727) that could be associated with the multitarget functions of NTZ, such as inhibition of the mTOR pathway.^30,31 Time-course studies on STAT3 phosphorylation after NTZ treatments at lower concentrations (20 or 40 μM) showed similar results (Figure S2). To further confirm that NTZ can block STAT3 activation after a longer treatment, the p-STAT3 (Tyr705), p-STAT3 (Ser727), and total STAT3 levels were measured after a 24 h interval. As expected, the levels of both phospho sites were reduced in a dose-dependent manner while the total STAT3 level remained stable (Figure 2C). In addition, NTZ displayed no obvious inhibition against STAT5, a STAT3 homologue, indicating potential selectivity for STAT3 (Figure S3).

Figure 2.

HeLa cells were treated with (A, C) the indicated doses of NTZ for 1 h or (C) 24 h or (B) with 80 μM NTZ for the indicated times, followed by Western blot assay with specific antibodies using total cell lysates.

To determine whether the suppression of phospho-Tyr705 came from direct inhibition of STAT3 or its upstream tyrosine kinases, we also measured the levels of both phospho- and total JAK2 and Src kinases, which are two prime direct activators of STAT3.^32,33 The results showed that after 1 or 24 h of treatment, NTZ did not affect the protein levels of Src and JAK2 or their phospho forms, including p-Src (Tyr416), p-Src (Tyr527), and p-JAK2 (Tyr1007/1008) (Figure S4). Considering that many kinds of direct SH2 binders that inhibit phosphorylation of Tyr705 have been reported to date and interfere with the dimerization of STAT3 in cells, we measured the direct binding of NTZ with the SH2 domain through a fluorescence polarization binding assay.³⁴ However, neither NTZ nor its active metabolite tizoxanide (TIZ) could directly bind to the SH2 binding site of STAT3, as they failed to interrupt the interaction of STAT3 protein with fluorescently labeled SH2 peptide (Figure S5). We also measured its direct binding with the DNA domain using an electrophoretic mobility shift assay (EMSA) and found that NTZ could inhibit the binding of STAT3 to its consensus DNA elements in both cell lysates and living cells (Figure S6). It was interesting to see a DNA domain binder affect the phosphorylation of Tyr705. In fact, some inhibitors targeting the STAT3 DNA binding domain have been reported to downregulate the p-STAT3 (Tyr705) level by increasing the susceptibility of phosphorylated STAT3 to phosphatase activity after blocking the STAT3–DNA interaction.³⁵ We therefore treated HeLa cells with NTZ in the presence or absence of sodium orthovanadate (Na₃VO₄), which is a phosphatase inhibitor, and found that the p-STAT3 (Tyr705) level in the cells pretreated with 25 μM Na₃VO₄ for 15 min prior to NTZ treatment was similar to that in non-Na₃VO₄-treated cells (Figure S7). These data suggested that as a DNA domain binder, NTZ could inhibit STAT3 phosphorylation by facilitating dephosphorylation of STAT3.

Because NTZ has been observed to effectively inhibit STAT3 signaling transduction, we measured the target protein levels of downstream cyclin D1, c-Myc, and survivin, which are supposed to be downregulated in response to STAT3 inhibition, to further validate the inhibitory activity of NTZ on the STAT3 pathway. The immunoblots demonstrated that NTZ effectively reduced the cyclin D1, c-Myc, and survivin levels in dose-dependent manner (Figure 3A). Accordingly, flow cytometry analysis also revealed that NTZ dose-dependently induced G0/G1-phase cell-cycle arrest and promoted apoptosis in HeLa cells (Figure 3B,C). The cell viability was inhibited by NTZ with an IC₅₀ value of 35.0 ± 0.1 μM (Figure 3D and Table 1). In summary, by inhibiting the phosphorylation at the two important sites of STAT3, NTZ blocks the transcriptional activity of STAT3 and triggers cell death in cancer cells.

Figure 3.

(A) HeLa cells were treated with the indicated concentrations of NTZ for 24 h, followed by Western blot analysis with specific antibodies using total cell lysates. (B, C) Percentages of cells in different proliferative or apoptotic states measured by flow cytometric analysis. DNA content was analyzed by propidium iodide (PI) staining in a cell-cycle assay, and the cell apoptotic state was analyzed by double-staining with Annexin V-FITC and PI. (D) Dose–response curve of NTZ against growth of HeLa cells measured by MTT assay after treatment for 48 h. Data are representative of three independent experiments.

Table 1. Structures and STAT3 Pathway Inhibition Potencies of Target Thiazolides.

^aNot tested.

^bSelectivity index: % inhibition of the STAT3 pathway divided by % inhibition of cell viability for HEK-Blue IL-6 cells. All of the values are averages of three independent experiments.

Considering its high safety and tolerability as an approved pediatric drug, NTZ could serve as a good lead compound for the development of antitumor agents. Thus, we designed and synthesized a series of NTZ derivatives to validate STAT3 as a target of the thiazolide scaffold and explore the preliminary structure–activity relationship on the phenyl ring, keeping the 5-nitrothiazole-2-amine moiety unchanged because the existence of 5′-nitro or halo substitution has been reported to be critical for the antitumor activity.³⁶ The detailed synthesis can be found in the Supporting Information. The STAT3 inhibitory activities of the prepared compounds were evaluated though the HEK-Blue IL-6 cell model described above. The IC₅₀ values were measured for those with inhibitory rates higher than 50%, while the cytotoxicity was relatively low as assessed by the MTT assay (selectivity index (SI) > 2).

As shown in Table 1, NTZ and TIZ exhibited similar potencies against the STAT3 pathway indicating that 2-OH was a sufficient substitution to inhibit STAT3 activity. When the 2-OH was moved to the 3- or 4-position, the activity was significantly decreased (1, 2). Attempts involving removal of the hydroxyl group (3), substitution with a methyl group (4), or introduction of an electron-withdrawing group such as F, Cl, Br, or nitro at the 2-position (5–8) also led to reduced activity. In order to identify whether there was other suitable space for further modifications on the benzoyl moiety, a methyl group was introduced at position 3, 4, or 5 while the 2-OH was retained (9–11). Encouragingly, the results showed that additional substitutions were beneficial for the activity, and the 4-methyl derivative 10 showed 4-fold higher activity than TIZ (10: IC₅₀ = 2.3 ± 0.08 μM; Table 1). However, introduction of a hydroxyl group instead of a methyl group at the 4-position eliminated the activity (10 vs 12).

Inspired by the above results, we subsequently focused on modifications at the 3- or 4-position. We found that removing the −OH while introducing −Cl at the 3- or 4-position generated 3–6-fold increased activity, as in 4-Cl derivative 14 (IC₅₀ = 1.5 ± 0.3 μM; Table 1). Electron-donating groups (2, OH; 17, OCH₃), bulky groups (18, 4-methylpiperazin-1-yl; 19, morpholino; 20, tert-butyl; 23, N-phenylsulfamoyl), or strongly polar groups (21, methylsulfonyl; 22, sulfamoyl) at the 4-position were all harmful to the activity. By contrast, small lipophilic substitutions at the 4-position, such as −Cl, −CF₃, or −N₃ (14–16), generated significantly improved potency. In particular, 15 exhibited a submicromolar IC₅₀ (0.7 ± 0.1 μM) and lower cytotoxicity (Table 1). The 3,4-dichloro analogue 24 was more potent than TIZ (IC₅₀ = 0.8 ± 0.04 μM; Table 1). However, the 3,4-dihydroxy analogue 25 was not active. In addition, the effect of amide −NH on the activity seemed indispensible because replacing the hydrogen atom with propynyl group resulted in a total loss of STAT3 activity (15 vs 26).

To further investigate the antitumor potency of the compounds with STAT3 pathway inhibitory activities higher than those of NTZ and TIZ, the antiproliferative activities were evaluated in HeLa cells. A significant correlation between the antitumor activity and the individual STAT3 inhibitory activity was observed, with a Pearson correlation coefficient of better than 0.99 (Figure S8). All of these compounds demonstrated satisfactory safety, as none of them induced cell death higher than 40% at a concentration as high as 50 μM in human normal embryonic kidney HEK 293T cells (Table 2). Notably, 15 and 24 exhibited even better activities than the positive control WP1066 in both STAT3 pathway activation and HeLa cell proliferation, while the cytotoxicities were much lower, indicating good potential of these compounds as antitumor agents. We selected 15 and measured its antiproliferative activities against diverse types of cancer cells (Table 3), among which PC3 prostate cancer and HT29 colon adenocarcinoma cells possess only low levels of activated STAT3 (Figure S9). Compound 15 exhibited much improved efficacies compared with NTZ in most of the cell lines, except for PC3 and HT29. These data suggest again that STAT3 inhibition plays an important role in the antitumor function of thiazolides.

Table 2. Growth-Inhibitory Activities of Thiazolides in the HEK 293T and HeLa Cell Lines^a.

			HEK 293T cell growth inhibition % (48 h)
compd	IC50 (μM, 20 h) STAT3 pathway in HEK-Blue IL-6 cells	IC50 (μM, 48 h) HeLa cell growth	50 μM	10 μM
NTZ	9.8 ± 0.2	35.0 ± 0.1	24.4 ± 1.0	–3.3 ± 4.6
TIZ	8.9 ± 0.3	29.8 ± 1.5	21.6 ± 3.8	–5.7 ± 4.3
9	4.6 ± 0.3	20.5 ± 4.4	21.2 ± 1.8	–1.0 ± 5.8
10	2.3 ± 0.08	16.4 ± 1.5	24.9 ± 03	11.0 ± 1.8
11	3.6 ± 0.3	17.7 ± 6.4	25.8 ± 4.1	–4.3 ± 2.9
13	3.0 ± 1.1	11.1 ± 1.7	–0.2 ± 3.4	–5.5 ± 3.3
14	1.5 ± 0.3	7.9 ± 2.1	21.0 ± 2.4	8.5 ± 3.6
15	0.7 ± 0.1	2.7 ± 1.1	38.3 ± 0.9	12.8 ± 2.7
16	2.5 ± 0.5	9.9 ± 1.7	30.8 ± 3.0	5.1 ± 3.0
24	0.8 ± 0.04	3.3 ± 0.8	19.1 ± 3.7	8.0 ± 4.6
WP1066	2.5 ± 0.7	4.2 ± 0.2	97.2 ± 0.2	79.9 ± 6.8

^aAll of the values are averages of three independent experiments.

Table 3. Growth-Inhibitory Activities of NTZ and 15 against Diverse Cancer Cell Lines^a.

	IC50 (μM, 48 h)
compd	HeLa	Caco-2	A549	A375	U87MG	HL60	PC3	HT29
NTZ	35.0 ± 0.1	26.8 ± 2.7	>50	31.8 ± 3.3	30.0 ± 3.5	20.1 ± 2.5	44.7 ± 3.0	>50
15	1.8 ± 0.04	1.8 ± 0.2	3.0 ± 0.4	2.8 ± 0.5	2.3 ± 0.9	1.2 ± 0.08	>10	>10

^aAll of the values are averages of three independent experiments.

We further analyzed the effect of 15 on STAT3 activation and transcription using Western blot assays. A time-course study revealed that 15 could significantly inhibit Tyr705 phosphorylation within 15 min and keep it at a very low level for a long term. There was no obvious change in Ser727 phosphorylation (Figure 4A). 15 dose-dependently decreased the protein level of p-STAT3 (Y705), whereas the p-STAT3 (S727) protein level remained stable after incubation for 24 h (Figure 4B). Additionally, 15 did not affect the total STAT3 protein level in either a time- or dose-dependent manner (Figure 4A,B). EMSA analysis revealed that similar to NTZ, 15 could also significantly block the interaction between STAT3 and DNA at both the molecular and cellular levels (Figure S6), facilitating the dephosphorylation of p-STAT3 (Y705) (Figure S7). The transcriptional activity of STAT3 was negatively regulated, resulting in an obvious decrease of downstream target proteins, as indicated by cyclin D1 and survivin (Figure 4C).

Figure 4.

Negative regulation of the transcriptional activity of STAT3. HeLa cells were treated with (A) 10 μM 15 for indicated times or (B, C) the indicated doses of 15 for 24 h, followed by Western blot assay with specific antibodies using total cell lysates.

As an ester prodrug intended to treat intestinal pathogens, NTZ can be partially taken up though passive absorption in the gastrointestinal tract, but its oral bioavailability is poor.³⁷ Upon oral absorption, NTZ is immediately deacetylated to give TIZ, which is subsequently metabolized as tizoxanide–glucuronide in the liver and rapidly eliminated via urine.³⁸ This pharmacokinetic property is perfect for an anthelmintic drug but much less desirable for a systemic antitumor agent. Finally, we selected 15 and further compared the in vivo pharmacokinetic properties in rats with those of NTZ. The plasma concentration–time curves are shown in Figure S10, and the key pharmacokinetic parameters are summarized in Table 4. After a single dose of oral administration, NTZ displayed a high clearance rate (6.3 L h^–1 kg^–1) and low absolute bioavailability (5.7%). By comparison, 15 exhibited more desirable pharmacokinetic properties, with a significantly longer half-life for elimination (t_1/2β) (11.1 vs 0.8 h), greater absolute bioavailability (F) (87.4% vs 5.7%), and higher maximum plasma concentration (C_max) (20.7 vs 1.0 mg/L).

Table 4. Pharmacokinetic Parameters of NTZ and 15 (n = 3 Rats/Group).

administration	t1/2α (h)a	t1/2β (h)b	Vd/F (L/kg)c	CL/F (L h–1 kg–1)d	AUC0–∞ (mg L–1 h–1)e	Tmax (h)f	Cmax (mg/L)g	F (%)h
NTZ (ig, 25 mg/kg)	0.6	0.8	0.	6.3	3.2	2.0	1.0	5.7
NTZ (iv 5 mg/kg)	0.05	0.3	0.08	0. 5	11.2	–	2.8	–
15 (ig, 25 mg/kg)	1.9	11.1	0.5	0.05	390.7	4.0	20.7	87.4
15 (iv, 5 mg/kg)	0.06	2.3	0.5	0.22	89.4	–	28.1	–

^at_1/2α: half-life for the distribution phase.

^bt_1/2β: half-life for the elimination phase.

^cV_d/F: apparent volume of distribution.

^dCL/F: apparent volume of the central compartment cleared of drug per unit time.

^eAUC_0–∞: area under the concentration–time curve.

^fT_max: time to peak concentration.

^gC_max: maximum plasma concentration.

^hF: absolute bioavailability.

In conclusion, we report that the FDA-approved antiparasitic drug NTZ can effectively inhibit the STAT3 pathway independent of interfering with the activities of the upstream kinases JAK2 and Src. In contrast to most reported SH2 binders, NTZ relies more on blocking the STAT3–DNA interaction. Furthermore, we performed preliminary structural deviations and structure–activity relationship analysis toward improved inhibitory activities against STAT3 pathway and identified eight thiazolides with potencies greater than that of NTZ. These derivatives exhibited increased antitumor potencies that were closely associated with their STAT3 inhibitory activities. Compared with WP1066, the optimized derivative 15 showed better inhibitory activity against both the STAT3 pathway and cancer cell proliferation, while its cytotoxicity was significantly lower. In addition, 15 exhibited greatly improved pharmacokinetic parameters compared with NTZ and showed broad-spectrum antitumor activities in multiple cancer cell lines, which indicates that 15 deserves further in vivo evaluation as an antitumor agent and that thiazolides represent a good scaffold for novel STAT3 inhibitor discovery.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.0c00544.

• Typical experimental procedures; supporting figures (Figures S1–S10) and full Western blots (Figures S11–S18); synthesis and characterization of target compounds; ¹H and ¹³C NMR and HRMS spectra (PDF)

Author Contributions

The manuscript was written through contributions of all authors.

The authors declare no competing financial interest.