Discovery, synthesis and biological evaluation of a series of N-(phenylcarbamothioyl)-2-napthamides as inhibitors of Claudin-1

Abstract

Colorectal cancer (CRC) remains a leading cause of cancer-related deaths worldwide, despite advancements in diagnosis. The main reason for this is that many newly diagnosed CRC patients will suffer from metastasis to other organs. Thus, the development of new therapies is of critical importance. Claudin-1 protein is a component of tight junctions in epithelial cells, including those found in the lining of the colon. It plays a critical role in the formation and maintenance of tight junctions, which are essential for regulating the passage of molecules between cells. In CRC, claudin-1 is often overexpressed, leading to an increase in cell adhesion, which can contribute to the development and progression of the disease. Studies show that high levels of claudin-1 are associated with poor prognosis in CRC patients and targeting claudin-1 may have therapeutic potential for the treatment of CRC. Previously, we have identified a small molecule that inhibits claudin-1 dependent CRC progression. Reported herein are our lead optimization efforts around this scaffold to identify the key SAR components and the discovery of a key new compound that exhibits enhanced potency in SW620 cells.

Graphical Abstract

Introduction:

Claudin-1 is a protein that plays a crucial role in the formation and maintenance of tight junctions, which are specialized structures that seal the gaps between adjacent cells in various tissues, including the lining of the colon.^1–3 Tight junctions help regulate the passage of molecules and ions across tissues, and they also play a role in the maintenance of tissue architecture and homeostasis.⁴ Recent research has suggested that claudin-1 may be involved in the development and progression of colon cancer.⁵ Several studies have shown that claudin-1 is overexpressed in colon cancer cells, and that this overexpression is associated with increased tumor growth, invasion, and metastasis.^6–8 For example, a recent study found that claudin-1 overexpression was associated with increased tumor size, lymph node involvement, and advanced tumor stage in patients with colon cancer.^{5, 6} Other studies have suggested that claudin-1 may play a role in the epithelial-mesenchymal transition (EMT), a process by which cancer cells acquire a more invasive phenotype and are better able to spread to other parts of the body.⁹ However, the exact mechanisms by which claudin-1 contributes to colon cancer development and progression are still not fully understood, and further research is needed to clarify the role of this protein in the disease.

Although the crystal structure of claudin-1 is unknown, we utilized YASARA, an in silico protein folding program, to identify a 3D structure which we then modeled using Molegro Virtual Docker (MVD).¹⁰ The model displayed putative cavities in the claudin-1 protein that could be targeted by small molecule inhibitors. Next, we screened the University of Nebraska Medical Center small molecule library (100K) and identified ten inhibitors that targeted the c-terminal domain of claudin-1 (the most accessible pocket). The candidates for further testing were selected based on the best binding energies.¹¹ These compounds were screened using SW480^claudin−1, SW620 and control cells and claudin-1 deficient HCT116 and SW620^shRNAcld−1 were used as negative controls. Resistance to anoikis (in vitro model for metastasis) and cell invasion was used as the biological output in testing the inhibitors, and we identified compound 1 as the most effective.^{12, 13} 1 mediated an increase in resistance to anoikis and cell invasion while HCT116 cells remained unaffected.¹¹ Lastly, compound 1 showed significant growth inhibition of xenograft tumors from SW480^claudin−1 cells in mice.¹¹ In this paper we report the structure-activity relationship studies around 1 and the discovery of a new compound as a next-generation claudin-1 inhibitor.

We have previously disclosed compounds 10 (KVA-E-25B) and 28 (PDS-0330) as inhibitors of claudin-1 and showed that 28 exhibited antitumor and chemosensitizer activities in an in vivo model of colorectal cancer.¹¹ As part of this effort, we undertook a medicinal chemistry campaign to evaluate the pharmacophore associated with the N-(phenylcarbamothioyl)-2-napthamides scaffold (Figure 1). The compound has several distinct areas for medicinal chemistry intervention starting with the naphthalene moiety (blue) and the 2-substituent. Next, the central phenyl ring (green) was replaced with pyridine and the right-hand benzimidazole (red) was replaced with a variety of heterocyclic groups. Lastly, we evaluated whether the acylthiourea was required by replacing it with amides and acylated ureas.

Figure 1.

SAR evaluation of the N-(phenylcarbamothioyl)-2-napthamides scaffold. Areas for SAR exploration are highlighted in the colored boxes and blue arrow.

Chemistry.

The synthesis of the final targets is shown in Scheme 1. The acylthioureas were synthesized following known procedures starting with the commercially available carboxylic acids. The acids, 1, were converted to the acid chlorides (oxalyl chloride, DCE, rt or thionyl chloride, acetonitrile, 70 °C) and the resulting acid chlorides, 2, were directly converted to the acylthioureas via reaction with ammonium thiocyanate, followed by the reaction with an appropriate aniline to yield the desired final targets, 10–37, 44–46 (NH₄SCN, 60 °C, ArNH₂).¹⁴ The ureas or thioureas, 4, (or cyclic acylated thioureas, 5) were synthesized by reaction of the aniline (or dihydroisoquinolinone) to yield the desired target compounds, 38–39 and 40–43. For the aniline compounds that were not commercially available we followed a procedure from Lam et al. utilizing an aromatic trifluoromethyl group as a synthetic handle to form the heterocycle.¹⁵ Thus, the trifluoroaromatic aniline, 7, was reacted with the 1,2-aminophenol, 1,2-aminothiophenol, or 1,2-diamine to yield the NH-containing heterocycles, 9 (1N NaOH, 60 – 90 °C). The procedure proceeded smoothly to afford aromatic anilines that were not previously accessible.

Scheme 1.

Synthetic procedures for the Claudin-1 inhibitors.

To evaluate the SAR around the newly synthesized compounds, we tested them in a cell-based assay for cell survival against the SW620 colorectal cancer cells which overexpress claudin-1 (Figure 2 and Table 1).¹¹ Our initial SAR survey was done at 25 μM and then promising compounds were moved to a dose response in the same cell line. As shown in Figure 1, the molecule affords multiple areas of SAR modification and we wanted to independently evaluate these areas to better understand their impact on the cell assay. The SAR campaign started with the right-hand heteroaryl section (compounds 1–18), using previously disclosed 10 (KVA-E-25B) as the comparator compound. Most of the analogs were not active, including the benzoxazole, 11 (KVA-E-25C), and benzothiazole, 13 (TMW-I-40), as well as the pyridine for phenyl replacements (12 (TMW-I-18) and 14 (TMW-I-39)). However, we did find success when truncating the right-hand heterocycle (compounds 15 (VM-A-156A) and 18 (VM-A-155B)); although, other five-membered heteroaryl groups were inactive, even the regioisomer oxadiazole 19 (VM-A-157A). Removing the thiourea (20 (VM-A-155A) and 22 (VM-A-162B)) proved unproductive, but replacing the thiourea with urea was successful in the direct comparator, 23 (VM-A-176), to 10. Next, our efforts moved to the left-hand portion of the molecule (naphthalene). Unfortunately, modification of this portion of the molecule proved to be unproductive as the only active compound was 28 (PDS-0330), which we previously identified. Changes to the methoxy group (isopropoxy, ethyl), adding a methoxy and moving to the quinoline resulted in inactive compounds, as was the dihydroquinolinone. Lastly, replacing the methoxynaphthalene with methoxyphenyl was not tolerated nor was removing the urea (as shown above as well).

Figure 2.

% Cell survival at 25 μM in SW620 cells.

Table 1.

SAR evaluation of synthesized compounds.

Cmpd	Structure	% inhibition at 25 uM	Cmpd	Structure	% inhibition at 25 uM
10 KVA-E-25B		68.4	19 VM-A-157A		18.4
11 KVA-E-25C		57.6	20 VM-A-155A		42.3
12 TMW-I-18		0	21 VM-A-162A		29.0
13 TMW-I-40		7.1	22 VM-A-162B		17.6
14 TMW-I-39		71.9	23 VM-A-176		90.8
15 VM-A-156A		89.9	24 VM-A-178		0
16 VM-A-156B		34.8	25 VM-A-180		0
17 KAT-I-131		33.1	26 VM-A-179		23.8
18 VM-A-155B		86.5	27 VM-A-177		42.6
28 PDS-0330		76.4	39 KAT-I-140A		0.9
29 KVA-E-23B		12.7	40 KVA-E-19A		23.8
30 TMW-I-26		0	41 KVA-E-19B		19.9
31 TMW-I-41		13.0	42 KVA-E-20A		13.2
32 TMW-I-30		17.2	43 KVA-E-20B		29.2
33 PS-I-5		37.6	44 KVA-E-22A		17.2
34 PS-I-14		15.8	45 KVA-E-22B		16.9
35 KAT-I-135		0	46 PS-I-18		0
36 KAT-I-137A		0	47 VM-A-104		44.8
37 KAT-I-137B		1.4	48 VM-A-109		27.8
38 KAT-I-140C		0

Having identified three compounds with significant activity at 25 μM, we wanted to further evaluate these compounds in a dose response assay, in addition to evaluating the compounds against normal IEC-6 cells to test for general cytotoxicity (Figure 3).¹¹ Again, 28 (PDS-0330) was used as a comparator as we have shown this compound to be active in an in vivo mouse model.¹¹ In SW620 cells, 28 showed a dose response curve with <50% cell survival starting at 12.5 μM. The new compounds each showed cell survival <15–20% at 6.25 μM, displaying better potency than 28. We also evaluated the compounds against the IEC-6 cell line as a control for general cytotoxicity and 15 (VM-A-156A) and 23 (VM-A-176) showed cytotoxicity at 12.5 and 6.24 μM, respectively, and these compounds were not pursued further.

Figure 3.

Dose response assay in SW620 and IEC-6 cells of select compounds.

To further assess this N-(phenylcarbamothioyl)-2-napthamide scaffold to determine their drug-like properties, we profiled select compounds in in vitro DMPK assays to assess their human and mouse liver microsomal intrinsic clearance, plasma protein binding and cytochrome P450 (CYP) inhibition (Table 2).^{16, 17} Our previously reported compounds, 10 and 28, displayed moderate stability in both human and mouse liver microsomes and were highly protein bound in both species (%f_u <0.5%). In fact, all compounds tested displayed very high plasma protein binding in both species; and had poor reproducibility in the assays, probably due to the high binding as they were stable in plasma. As both compounds showed similar stability in microsomes, it did not appear that the presence of the methoxy group in 10, a common metabolic liability, was contributing to any instability. In fact, moving to the benzoxazole, 11, provided a stable compound in both human and mouse liver microsomes. Moving to the right-hand truncated heterocycles introduced instability in human liver microsomes (15 and 16), but these were moderately stable in mice. The oxadiazole compounds (18 and 19) regained the stability in human microsomes while retaining the moderate stability in mice. The urea compounds, 23 and 27, showed similar stability to the thiourea counterparts. We tested our two most active compounds, 15 and 18, against a panel of CYP enzymes (3A4, 2D6, and 2C19) and both compounds were inactive (<50% inhibition at 10 μM).

Table 2.

In vitro PK parameters of selected compounds.

	Intrinsinc Clearance (mL/min/kg)a,c				Plasma Protein Binding (%fu)b,c		CYP Inhibition (% at 10 μM)c
Cmpd	hCLINT	hCLHEP	mCLINT	mCLHEP	hPPB	mPPB	CYP3A4	CYP2D6	CYP2C19
10 KVA-E-25B	27.8	11.7	42.7	29.0	<0.03	<0.01	NDd
11 KVA-E-25C	<20	<10	<20	<16	*	<0.03
15 VM-A-156A	125.0	17.3	57.8	35.2	*	*	28	36	43
16 VM-A-156B	201.5	18.3	37.7	26.6	*	2.2	NDd
18 VM-A-155B	<20	<10	38.2	26.9	*	*	16	19	37
19 VM-A-157A	<20	<10	46.2	30.5	*	<0.03	NDd
23 VM-A-176	67.8	15.5	235.5	65.2	<0.3	<0.3
27 VM-A-177	92.8	16.5	43.5	29.3	<0.3	<0.3
28 KVA-E-23A	32.4	12.4	53.4	33.5	<0.04	<0.03
29 KVA-E-23B	86.3	16.3	364.6	72.2	*	*
38 KAT-I-140C	<20	<10	139.1	54.7	<0.03	*

^aPredicted hepatic clearance based on intrinsic clearance in mouse and human liver microsomes using the well-stirred organ CL model (binding terms excluded).

^bf_u = fraction unbound.

^cIn vitro DMPK studies performed at Q2 Solutions, Indianapolis, IN.

^dND = not determined.

^*Poor reproducibility.

Based on the totality of the properties of the evaluated compounds, we chose to progress 18 into a discrete mouse in vivo PK study (IP dosing, 5 mg/kg) (Table 3). In addition, we chose to evaluate 15 in an in vivo PK study, even though it was not progressed due to its cytotoxicity, we felt this may provide information that would be valuable for future SAR studies. The study was done to determine the plasma concentrations over a 24 h period to inform any future in vivo animal studies. Both 15 and 18 displayed moderate to good plasma half-life (t_1/2 = 2.5 h and 3.5 h, respectively). Compound 18 had nearly twice the AUC concentrations and T_max versus 15. Overall, 18 displayed favorable in vivo PK conditions to be used in future animal studies of colon cancer.

Table 3.

In vivo mouse PK.

	15, VM-A-156A	18, VM-A-155B
In vivo PK parametersa,b
T1/2 (h)	2.49 ± 0.59	3.51 ± 0.51
MRT (h)	2.70 ± 0.06	5.30 ± 0.15
Tmax (h)	1.00 ± 0.00	2.33 ± 1.15
Cmax (ng/mL or ng/g)	5063±225	5100 ± 752
AUClnf (h*ng/mL)	15868±958	34111±3813

^aPK studies performed at Pharmaron, Inc. (Louisville, KY).

^bIP dosing (5 mg/kg); Formulation (10%DMSO/10% Cremphor EL/30% PEG400/50% water) in CD-1 mice; Dosed as a suspension; PK parameters were estimated by non-compartmental model using WinNonlin 8.3.

In conclusion, building on our recently described efforts at the discovery of 10 and 28, we have described the synthesis and biological evaluation of a series of N-(phenylcarbamothioyl)-2-napthamides as Claudin-1 inhibitors. We synthesized a variety of compounds with changes to each of the major areas of the molecule and have documented the activity against the SW620 cells. The data show steep SAR surrounding the molecule where the naphthalene, thiourea (or urea) are required for activity. In addition, the 2-position only tolerated methoxy or hydrogen. The right-hand heteroaryl portion provided some tolerance for smaller five-membered rings, however. We identified three compounds with increased activity against the original compounds (15, 18, and 23) with 18 being the compound with the best overall properties (activity, in vitro and in vivo PK). Further studies characterizing 18 in animal models of colorectal cancer and in progress and will be reported in due course.