Synthesis and Evaluation of Novel Erlotinib–NSAID Conjugates as More Comprehensive Anticancer Agents

Abstract

A series of novel anticancer agents were designed and synthesized based on coupling of different nonsteroidal anti-inflammatory drugs (NSAIDs) with the epidermal growth-factor receptor (EGFR) tyrosine kinase inhibitor, erlotinib. Both the antiproliferative and pharmacokinetic activity of the target compounds were evaluated using HCC827 and A431 tumor cell lines. Among the derivatives made, compounds 10a, 10c, and 21g showed superb potency, comparable to that of erlotinib. Furthermore, preliminary SAR analysis showed that when the NSAIDs were conjugated via linkage to C-6 OH versus linkage to C-7 OH of the quinazoline nucleus, superior anticancer activity was achieved. Finally, the in vitro pharmacokinetic profile of several conjugates demonstrated the desired dissociation kinetics as the coupled molecules were effectively hydrolyzed, releasing both erlotinib and the specific NSAID in a time-dependent manner. The conjugation strategy represents a unique and simplified approach toward combination therapy, particularly for the treatment of cancers where both EGFR overexpression and inflammation play a direct role in disease progression.

Lung cancer, in particular nonsmall cell lung carcinoma (NSCLC), is a disease associated with a high mortality rate due to its complexity and heterogeneity.¹⁻³ Although there has been significant progress made in its diagnosis, management, and treatment, the five year survival rate for this disease is still low.⁴ Chemotherapy is a primary option for patients with lung cancers like NSCLC and the use of epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors has become a frontline approach due to their demonstrable effectiveness in slowing the progression of the disease as well as increasing the overall survival rate of patients.⁵⁻⁷

However, in cancer treatment, using a single drug directed toward just one target/mechanism is usually inadequate and overtime disease management usually requires drug combinations targeting multiple cancer pathways to sustain efficacy.⁸⁻¹⁰ Choosing the right combination of chemotherapeutic agents is crucial for maximizing progression free survival while minimizing side effects.

Inflammation has received considerable attention for its role in the progression and aggressiveness of many cancers including NSCLC. Many studies have shown the efficacy of NSAIDs in animal models of cancer prevention, partly due their ability to block cyclooxygenase (COX) activity.¹¹⁻¹⁹ Mechanistically, cyclooxygenases including COX-1 and especially COX-2 have been shown to be important in many stages of oncogenesis. A case in point is the USA FDA approval of the NSAID, celecoxib (Celebrex, a selective COX-2 inhibitor), for use in patients with familial adenomatous polyposis. Also, many studies have shown the efficacy of COX-2 inhibitors in various animal models of cancer. Mechanistically, COX-2 is highly expressed in many tumor and stromal cells and PGE2, a major product of COX-mediated arachidonic acid metabolism, is believed to be a contributing component of angiogenesis in lung cancer. Therefore, a combination drug that targets both EGFR and cyclooxygenases may be a more comprehensive and effective treatment modality for the treatment of lung cancer.

In this study we present a series of esterase hydrolyzable NSAIDs conjugated with erlotinib, which may have potential as a new class of chemotherapeutic agents that offers simplified dosing regimen with improved efficacy.

The preparation of conjugated NSAIDs with erlotinib 10a–f, 11a–g, and 21a–g is outlined in Schemes 1 and 2. Compound 7 was obtained from commercially available 2-amino-4,5-dimethoxybenzoate (1) and elaborated according to the procedure previously described.^20,21 Then the 6-OH of compound 7 was reacted with 1-bromo-2-methoxyethane in the presence of K₂CO₃ to generate 8. The desired conjugated derivatives 10a–f were obtained by reacting 8 with 9a–f, which had been prepared by esterification the corresponding acid with 2-bromoethanol in the presence of SOCl₂. Analogues 11a–g were prepared by a similar experimental procedure.

Scheme 1. Synthesis of Compounds 10a–f and 11a–g.

Reagents and conditions: (a) HCO₂NH₄, HCONH₂, 140 °C, 24 h; (b) HBr, Ac₂O; (c) Ac₂O, py; (d) POCl₃, CHCl₃; (e) 3-ethynylaniline, i-PrOH, reflux; (f) NH₄OH; (g) 1-bromo-2-methoxyethane (1.0 equiv), K₂CO₃, DMF, 20 °C; (h) 9a–f, K₂CO₃, DMF; (i) 2-bromoethanol, SOCl₂.

Scheme 2. Synthesis of Compounds 21a–g.

Reagents and conditions: (a) 1-bromo-2-methoxyethane, NaH, DMF; (b) 2-bromoethyl acetate, K₂CO₃, DMF; (c) conc. HNO₃, conc. H₂SO₄, HOAc; (d) Fe, HOAc; (e) HCOONH₄, HCONH₂; (f) POCl₃, CHCl₃; (g) 3-ethynylaniline, i-PrOH, reflux; (h) NaOH, MeOH; (i) COCl₂, DCM; (j) 9a–f, Et₃N, DCM.

Ethyl 3,4-dihydroxybenzoate was selectively alkylated with 1-bromo-2-methoxyethane to afford 12, which was then converted to compound 13 in the presence of 2-bromoethyl acetate and K₂CO₃. Nitration of 13 with concentrated HNO₃ and H₂SO₄ in acetic acid gave 14, which was converted into aniline 15 by reduction with iron in acetic acid. Quinazolin-4(3H)-one 16 was prepared by condensation of 15 and HCOONH₄ in HCONH₂. Treatment of 16 with POCl₃ in CHCl₃ gave 4-chloroquinazoline 17, which was treated with 3-ethynylaniline in isopropyl alcohol to afford 4-anilinoquinazoline derivative 18. Hydrolysis of the acetyl group of compound 18 was performed using sodium hydroxide in methanol to provide the key intermediate 19. The desired conjugated compounds 21a–g were prepared by reacting 19 with the corresponding NSAID acyl chloride using Et₃N as the base in DCM.

To assess the potency of the conjugated compounds versus the individual noncoupled molecules, selected compounds were tested against EGFR tyrosine kinase, COX-1 and COX-2 enzymes (Table 1). As expected, the conjugated compounds showed less activity in their coupled state compared to the corresponding unconjugated molecules. The noncoupled compounds effectively blocked their target enzymes with IC₅₀ values consistent with previous reports.²²⁻²⁶ These data suggest that the conjugated compounds are likely acting as pro-drugs for their parent molecules. Release of the parent molecules from these pro-drugs may be advantageous from a safety perspective as the plasma levels of fully active drugs may be more precisely managed using this strategy. In addition, it has been reported that esterases have elevated activity in tumors suggesting that activation of the drug conjugates may be enhanced at sites of disease relative to normal tissues.^27,28

Table 1. Inhibition of EGFR, COX-1, and COX-2.

enzyme inhibition IC50 value (μM)
compd	EGFR	COX-1	COX-2
erlotinib	0.0005 ± 0.0004	>10	>10
aspirin	>10	>10	>10
ibuprofen	>10	>10	1.1 ± 2.27
sulindac	>10	16 ± 10.975	8.8 ± 2.72
naproxen	>10	0.18 ± 0.82	0.28 ± 0.21
indomethacin	>10	0.11 ± 0.05	0.41 ± 0.06
ketoprofen	>10	0.067 ± 0.05	0.061 ± 0.05
10a	0.005 ± 0.0006	>10	>10
10b	0.009 ± 0.0143	44.67 ± 20.61	>10
10c	0.031 ± 0.0034	>10	>10
10d	0.072 ± 0.0200	24.48 ± 14.71	>10
10e	0.027 ± 0.0015	11.34 ± 21.40	>10
10f	0.032 ± 0.0037	>10	>10
11a	0.034 ± 0.0068	39.57 ± 12.97	>10
11b	>10	>10	>10
11c	0.027 ± 0.0019	>10	>10
11d	0.39 ± 0.0480	>10	>10
11e	0.033 ± 0.0028	31.38 ± 0.31	>10
11f	0.88 ± 0.2820	>10	>10
GJT-0383	0.001 ± 0.0003	>10	>10

To assess the efficacy of the conjugated compounds as anticancer agents, a series of coupled molecules were evaluated for their antiproliferative activity against two tumor cell lines HCC827 and A431. As hoped, multiple conjugated drugs demonstrated excellent activity in blocking tumor cell proliferation (Table 2). Coupled compounds 10a, 10b, 10c, 10d, 10f, 21a, 21c, and 21g were particularly active at blocking proliferation of HCC827 cells. The efficacy potency of these compounds was comparable to erlotinib, which was used as a positive control. SAR analysis showed that NSAIDs conjugated via linkage to C-6 OH versus linkage to C-7 OH of the quinazoline nucleus possessed superior in vitro anticancer activity. It is important to point out that tumor cell-based assays are not adequate to assess the synergistic activity between erlotinib and the specific NSAID as cells grown under these conditions lack a significant inflammatory component. Future studies will focus on testing candidate conjugated drugs in animal models of cancer.

Table 2. Summary of the in Vitro Antiproliferative Activities of Conjugate Compounds against Two Human Cell Lines Including Epidermoid and Lung Adenocarcinoma.

	IC50 (μM)			IC50 (μM)
compd	HCC827	A431	compd	HCC827	A431
10a	0.1 ± 0.03	1.8 ± 0.49	10b	0.4 ± 0.23	2.2 ± 0.49
10c	0.3 ± 0.18	5.4 ± 1.83	10d	0.1 ± 0.04	3.2 ± 0.10
10e	0.9 ± 0.36	>10	10f	0.2 ± 0.05	>10
11a	0.9 ± 0.27	>10	11b	16.5 ± 18.35	NA
11c	3.5 ± 2.13	>10	11d	2.3 ± 0.32	NA
11e	1.6 ± 0.82	>10	11f	>10	NA
11g	2.1 ± 0.74	>10	21a	0.1 ± 0.04	>10
21b	>10	>10	21c	0.4 ± 0.29	4.8 ± 2.33
21d	0.5 ± 0.28	5.0 ± 0.21	21e	1.2 ± 0.46	>10
21g	0.3 ± 0.21	1.7 ± 1.35	erlotinib	0.029 ± 0.004	0.844 ± 0.02
GJT-0383	0.014 ± 0.003	1.27 ± 0.17

To make sure the drugs were acting on their respective targets, namely, EGFR tyrosine kinase and COX-2, as a representative example, cells and media were collected from 10d treated HCC827 cultures and analyzed for a reduction in both phosphorylation of EGFR and PGE2 levels by Western blot analysis and HPLC, respectively. As expected, a dose-dependent reduction of phosphorylation of EGFR (Figure 1) and PGE2 (IC₅₀ 23 μM; data not shown) was observed indicating that 10d is acting on its target/mechanism(s).

Figure 1.

Western blot analysis showing the effects of 10d at concentrations of 0 to 500 nM on the phosphorylation of EGFR in HCC827 cells.

To assess the time-dependent hydrolysis of the conjugated compounds and the corresponding dissociation of the parent molecules, a pharmacokinetic analysis was performed using compound 10d in cell culture. In the experiment, 10d was added to culture media and then allowed to incubate with HCC827 tumor cells for up to 24 h. During the incubation, conditioned medium was collected at various time points and assessed for the coupled compound 10d, and the parent molecules naproxen and erlotinib. In the case of erlotinib, following hydrolysis of compound 10d an additional hydrogen atom is left behind giving the molecule a slightly higher molecular mass and referred to as GJT-0383 in this study (Scheme 3). To make sure GJT-0383 is as potent as erlotinib, the compound was tested for activity against EGFR tyrosine kinase. Both GJT-0383 and erlotinib showed comparable IC₅₀ values (Table 1).

Scheme 3. Hydrolysis of Compound 10d.

As can be seen in Table 3, there is a time-dependent hydrolysis of 10d resulting in the release of both GJT-0383 and naproxen overtime. Within 24 h approximately 50% of compound 10d was hydrolyzed demonstrating that the conjugated drug can be processed in a cell based system. With the loss of compound 10d there is a corresponding increase in the concentrations of GJT-0383 and naproxen. At time 0 a small amount of GJT-0383 and naproxen was detected suggesting some hydrolysis during storage of 10d.

Table 3. Hydrolysis of Compound 10d in HCC827.

time (h)	10d (μg/L)	GJT-0383 (μg/L)	naproxen (μg/L)
0	3707 ± 4.714	105.0 ± 2.160	91.20 ± 0.8832
0.25	3567 ± 26.24	158.3 ± 4.110	108.4 ± 8.536
0.5	3427 ± 12.47	224.3 ± 4.922	128.7 ± 9.672
2	3233 ± 12.47	704.7 ± 3.682	293.0 ± 39.91
4	3150 ± 21.60	1277 ± 18.86	382.0 ± 29.34
8	2923 ± 33.00	2363 ± 9.428	849.0 ± 87.41
24	2137 ± 20.55	4397 ± 123.9	2017 ± 65.49

In the cell based experiments, active serum was used bringing up the possibility that esterases present in the serum are partly or mostly responsible for the hydrolysis of 10d overtime. In order to address this, 10d was incubated with both active versus heat inactivated whole serum and measuring the hydrolysis overtime. It was observed that 10d was effectively hydrolyzed overtime in the active serum, but in contrast, little or no hydrolysis of 10d occurred in the heat inactivated serum presumably due to the denaturing of the esterases contained in the serum (Figure 2). Based on these data it appears that a major proportion of the activation of 10d observed in HCC827 and A431 cells is mediated by serum esterases.

Figure 2.

Hydrolysis of 10d in active versus heat inactivated serum.

In this study we describe a new series of conjugated compounds that may be used as a potentially simplified approach for treating cancers where both EGFR overexpression and inflammation are involved. Future studies will focus on the in vivo efficacy of the more potent coupled compounds in animal models of colon and lung cancer where inflammation is believed to be a significant contributing factor to disease progression.

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmedchemlett.5b00286.

• Complete experimental details along with synthesis and characterization of Erlotinib–NSAID conjugates (PDF)

This work was supported in part by the Guangzhou Science & Technology Project (2011Y2-00026), Guangdong Science and Technology International Cooperation projects (Grant No. 2013B050800009) (to Y.Z.), CAS Key Technology Talent Program China (2013) (to Z.T.) and Guangdong Science and Technology Plan projects, China (Grant No. 2013B040200031) (to Z.T.).

The authors declare no competing financial interest.