Overcoming the imatinib-resistant BCR-ABL mutants with new ureidobenzothiazole chemotypes endowed with potent and broad-spectrum anticancer activity

Abstract The design of kinase inhibitors targeting the oncogenic kinase BCR-ABL constitutes a promising paradigm for treating chronic myeloid leukaemia (CML). Nevertheless, the efficacy of imatinib, the first FDA-approved targeted therapy for CML, is curbed by the emergence of resistance. Herein, we report the identification of the 2-methoxyphenyl ureidobenzothiazole AK-HW-90 (2b) as a potent pan-BCR-ABL inhibitor against imatinib-resistant mutants, particularly T315I. A concise array of six compounds 2a–f was designed based on our previously reported benzothiazole lead AKE-5l to improve its BCR-ABLT315I inhibitory activity. Replacing the 6-oxypicolinamide moiety of AKE-5l with o-methoxyphenyl and changing the propyl spacer with phenyl afforded 2a and AK-HW-90 (2b) with IC50 values of 2.0 and 0.65 nM against BCR-ABLT315I, respectively. AK-HW-90 showed superior anticancer potency to imatinib against multiple cancer cells (NCI), including leukaemia K-562. The obtained outcomes offer AK-HW-90 as a promising candidate for the treatment of CML and other types of cancer.


Introduction
Chronic myeloid leukaemia (CML) is a myeloproliferative neoplasm characterised by excessive production of immature white blood cells and constitutes approximately 15% of leukaemia cases in adults 1 . The fusion oncoprotein product of the Philadelphia chromosome (Ph), Break-point cluster region-Abelson (BCR-ABL), is the critical driver for the pathogenesis of CML and acute lymphoblastic leukaemia (Ph þ ALL) 2 . Therefore, the design of BCR-ABL kinase inhibitors offers a legitimate approach for the treatment of CML and ALL.
Imatinib (GleevecV R , (Figure 1)), a diarylamide-4-(pyridin-3-yl)pyrimidine conjugate, is the first FDA-approved BCR-ABL inhibitor for the treatment of CML patients 3 . Despite its initial significant response in most of the CML patients, the clinical efficacy of imatinib was hampered in about 40% of patients due to dose intolerance and the emergence of drug resistance 4 . Point mutations within the BCR-ABL kinase domain represent the most recurrent mechanism of imatinib resistance 5 , where they hinder its proper binding with the target 6 . Among more than 100 different mutations detected in CML patients, the gatekeeper T315I mutation is the most refractory form, which emerges in more than 20% of CML patients 7 . Second-generation BCR-ABL inhibitors, nilotinib (TasignaV R ), dasatinib (SprycelV R ), and bosutinib (BosulifV R ) ( Figure  1), have been approved to treat CML adult patients with acquired resistance to imatinib [8][9][10] . However, these second-line drugs failed to inhibit several imatinib-resistant mutations, including the T315I form 11 .
In 2012, the multikinase inhibitor ponatinib (IclusigV R , Figure 2) received FDA approval as the third-generation ABL inhibitor for the treatment of patients with CML and Ph þ ALL 12 . It demonstrated excellent potency against imatinib-clinically resistant ABL mutants, including T315I and other forms like Q252H and H396P 13,14 . However, the serious vascular/cardiotoxic effects associated with ponatinib restricted its use for these tumours harbouring T315I mutant. Ponatinib cardiotoxicity is thought to arise from its multikinase inhibitory mode and synchronous inhibition of kinases that are important for cardiovascular function 15 . Therefore, a number of structurally assorted BCR-ABL T315I inhibitors with improved selectivity/safety have been developed ( Figure 2). Olverembatinib (GZD824), a pyrazolopyridine-diarylamide conjugate, was developed from ponatinib as a potent BCR-ABL T315I inhibitor with a favourable safety profile 16 18 . In a previous study, our group reported a series of benzothiazoles featuring picolinamide moiety as BCR-ABL inhibitors 19 . In this set, compound AKE-5l ( Figure 3) was identified as the most potent BCR-ABL inhibitor with IC 50 values of 18.2 and 39.9 nM against wild BCR-ABL and its T315I mutant, respectively. Moreover, AKE-5l showed selective anti-leukemic activity against only the K-562 cell line out of 58 tested cancer cells. In continuation of our ongoing efforts to identify new chemical entities as potent anticancer kinase inhibitors, [20][21][22][23][24][25] we herein aimed at further optimisation of AKE-5l activity. Based on the exceptional potency of HS-438 towards BCR-ABL T315I (IC 50 ¼ 0.064 nM), it was evident that installing 2-methoxyphenyl motif directly at C-6 of benzothiazole is optimal for activity than 6-oxypicolinamide. Therefore, in the current study, we replaced the picolinamide ether of AKE-5l with 2-methoxyphenyl moiety, while conserving the ureidobenzothiazole scaffold (Figure 3). In addition, the aliphatic propyl spacer, which links ureidobenzothiazole and N-methylpiperazine in AKE-5l, was replaced by a phenyl ring to improve the anticancer spectrum while retaining the BCR-ABL inhibitory activity. The terminal phenyl was substituted with various hydrophilic and lipophilic moieties to underscore their impact on both kinase activity and  cellular potency. These two major structural modifications generated a concise library of six 1-(6-(2-methoxyphenyl)benzo[d]thiazol-2-yl)-3-phenylurea derivatives 2a-f. All target compounds were evaluated against both BCR-ABL WT and BCR-ABL T315I , and a representative derivative 2b was further profiled against an array of imatinib-resistant BCR-ABL mutants.

Experimental
The detailed experimental section is included in the supplemental material.

Chemistry
The synthetic outline for the preparation of the final compounds is illustrated in Scheme 1. The key starting material, 6-(2-methoxyphenyl)benzo[d]thiazol-2-amine (1), was obtained via Suzuki coupling of 2-amino-6-bromobenzothiazole with (2-methoxyphenyl) boronic acid using Pd(dppf)Cl 2 CH 2 Cl 2 as a catalyst under two different conditions. Initially, the reaction was conducted in 1,4dioxane:H 2 O, (3:1, v/v) at 95 0 C for 4 h using potassium carbonate as a base, and afforded compound 1 in 58% yield. Changing the solvent with 1,2-dimethoxyethane (DME):H 2 O (3:1, v/v), using sodium bicarbonate and running the reaction at 85 0 C for 2 h furnished 1 in higher yield (87%). Treatment of benzothiazol-2-yl amine 1 with 1,1 0 -carbonyldiimidazole (CDI) in DMF afforded the corresponding isocyanate, which was coupled with the proper aniline at 100 0 C to generate the target ureidobenzothiazoles.   >100000 a All compounds were tested in a 10-dose IC 50 mode with 3-fold serial dilution starting at 20 lM or 100 lM (for imatinib), and the reactions were conducted at 10 lM ATP. b Reported values. 19 In vitro kinase screening All final compounds 2a-f were tested against the native BCR-ABL WT along with its clinically imatinib-resistant T315I mutant at Reaction Biology Corporation (RBC, Malvern, PA, USA) 26 , using imatinib as a reference compound and the pan kinase inhibitor staurosporine as a positive control ( Table 1). As disclosed from the data, ureidobenzothiazoles 2a-c, with hydrophilic cyclic amine at the 4-position of the terminal phenyl ring, showed excellent inhibitory potency with IC 50 values spanning in the range of 0.701-2.16 nM and 0.651-8.11 nM against BCR-ABL WT and BCR-ABL T315I , respectively, being superior to the lead compound AKE-5l (BCR-ABL WT IC 50 ¼ 18.2 nM, BCR-ABL T315I IC 50 ¼ 39.9 nM). Such biochemical outcomes indicate that replacing the oxypicolinamide moiety of AKE-5l with o-methoxyphenyl, along with changing the propyl spacer into phenyl, resulted in improved BCR-ABL inhibitory activity. The existence of terminal nitrogen with small alkyl substituents, 2a and 2b, was found to be optimal for BCR-ABL inhibition. Compound 2b, with N-ethylpiperazine, exerted the best activity with sub-nanomolar IC 50 values of 0.701 and 0.651 nM against BCR-ABL WT and BCR-ABL T315I , respectively. It showed 2-3 folds better activity than its methylpiperazine congener 2a (BCR-ABL WT IC 50 ¼ 1.50 nM, BCR-ABL T315I IC 50 ¼ 2.03 nM). Replacing the terminal nitrogen of piperazines 2a and 2b with oxygen 2c retained the BCR-ABL WT activity, however, it led to reduction in BCR-ABL T315I inhibition (BCR-ABL T315I IC 50 ¼ 8.11 nM). Insertion of m-trifluoromethyl group on the distal phenyl while removal of the hydrophilic cyclic amines 2d resulted in a dramatic drop in potency (BCR-ABL WT IC 50 ¼ 681 nM, BCR-ABL T315I IC 50 ¼ 1240 nM). Such decline in activity was much more pronounced upon the introduction of an additional trifluoromethyl group at the 5-position, as in 2e (BCR-ABL WT IC 50 ¼ 4090 nM, BCR-ABL T315I IC 50 ¼ 9560 nM). In contrast, incorporation of morpholine methylene at the 4-position of 2d afforded 2f, which restored inhibitory potency with IC 50 values of 74.6 and 226 nM against BCR-ABL WT and BCR-ABL T315I , respectively. These findings point out the indispensable role of cyclic amines, particularly piperazine, in achieving sound BCR-ABL inhibition for this set of ureidobenzothiazoles. This importance might be attributed to the hydrophilic nature of cyclic amines, which offers a favourable orientation towards the solvent exposure region of BCR-ABL. Such a hypothesis could be emphasised by noticing the detrimental impact of the existence of m-trifluoromethyl group (2d and 2e) on the distal phenyl ring.
As an example of the piperazine ureidobenzothiazoles, compound 2b was further profiled against a panel of clinically imatinib-resistant BCR-ABL mutants, including the P-loop mutants E255K and Q252H, activation loop mutant H396P, catalytic segment mutant M351T, and mutant F317I in the ATP binding region ( Table 2). The obtained findings confirmed the potent ability of 2b to inhibit the activity of the clinically relevant BCR-ABL with single-digit nanomolar IC 50 values.
On the other hand, to gain certain insights about the kinase selectivity of this array of ureidobenzothiazoles, compounds 2a and 2b were tested against DDR1 ( Figure 4). Discoidin domain receptors (DDR1/2) are transmembrane receptor tyrosine kinases (RTKs), activated by fibrillar collagens 27 . DDR1/2 and BCR-ABL share almost 61% sequence similarities in their ATP binding domains. Therefore, several BCR-ABL inhibitors like imatinib, nilotinib, and dasatinib possess potent nanomolar DDR inhibitory   effects comparable to BCR-ABL 28 . Interestingly, compounds 2a and 2b showed 130 and 284 folds selectivity for BCR-ABL (IC 50 ¼ 1.5 and 0.7 nM, respectively) over DDR1 (IC 50 ¼ 195 and 199 nM, respectively). This observation points out the merit of these ureidobenzothiazoles over the known BCR-ABL inhibitors in achieving exceptional selectivity towards BCR-ABL.

In vitro cell-based evaluation of the antiproliferative activities
Assessment of the antiproliferative activity by MTT assay Motivated by the promising results of cell-free assay, we examined the antiproliferative activity of target compounds against the BCR-ABL overexpressing human leukaemia K-562 cells, along with L132 normal cell by MTT (3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide) assay (Table 3). Consistent with biochemical kinase findings, the most active ureidobenzothiazoles 2a-c displayed good anticancer activity against K-562 cell with GI 50 values of 3.39-4.52 mM. On the other hand, the m-trifluoromethylphenyl substituted members 2d-f exerted modest cellular potency (GI 50 > 10 mM). These cellular outcomes define the impact of BCR-ABL inhibition on the proliferation of K-562 cells. Moreover, compounds 2a-c showed weak cytotoxic effects over L132 normal cell, as evidenced by high GI 50 values (17.86-26.78 mM). In terms of selectivity index (SI), the top potent compounds 2a-c had reasonable SI values spanning in the range of 5.3-6.8, which indicate their selective antiproliferative activity towards cancer cell K-562.
Broader assessment of anticancer activity by sulforhodamine B (SRB) assay Single dose assay. To gain further insights about the anticancer profile of this new ureidobenzothiazole chemotypes, the structures of all final compounds were submitted to the National Cancer Institute (NCI, Developmental Therapeutics Program, Bethesda, MD, USA). Based on NCI selection criteria, compounds 2b (AK-HW-90), 2e, and 2f were chosen for broad screening against a library of 60 human cancer cell lines, representing nine types of blood and solid cancers. Employing the highly sensitive sulforhodamine B (SRB) assay 29 , compounds 2b, 2e, and 2f were tested initially at a single dose of 10 lM, and their detailed activities on the percentages growth of the NCI-60 cell lines were illustrated in Figure 5, and compared with that observed for the lead benzothiazole AKE-5l.
Interestingly, all three compounds 2b, 2e, and 2f showed superior anticancer potency (low % growth values) rather the lead ureidobenzothiazole AKE-5l. While AKE-5l showed narrow spectrum activity towards only K-562 cell line (% growth ¼ 10.22), the o-methoxyphenyl benzothiazoles 2b, 2e, and 2f exerted broad spectrum anticancer effect against multiple cancer cells ( Figure 5). Among these compounds, the most potent BCR-ABL inhibitor 2b elicited the best cellular activity, where it manifested remarkable lethal effects (minus % growth values) over 55 cancer cells. Since both lead AKE-5l and 2b (AK-HW-90) share N-alkylpiperazine and ureidobenzothiazoles as common structural features, it was evident that replacing oxypicolinamide moiety and propyl spacer of AKE-5l with o-methoxyphenyl and phenyl ring in 2b, respectively, led to dramatic change in the anticancer property from narrow into expansive spectrum covering diverse cancer cells. Next in the activity rank is the 3,5-bis(trifluoromethyl)phenyl derivative 2e, which showed cytotoxic rather cytostatic activity against 21 cancer cells. In contrast to 2b and 2e, compound 2f evinced reasonable cytostatic growth inhibitory activity (% growth less than 40) towards 14 cancer cells. Despite of the modest BCR-ABL inhibitory potency of 2e and 2f, they showed good anticancer activity over NCI-60 cell lines, particularly 2e, which points out the existence of other underlying mechanism(s), beside kinase inhibition, that contribute to the anticancer property of this novel chemotype.
Five dose assay. The promising anticancer activities of compounds 2b and 2e fulfilled the NCI criteria, therefore, they were advanced to five-dose testing mode to determine their GI 50 (the molar concentration causing 50% growth inhibition), TGI (the molar concentration producing 100% GI) and LC 50 (the molar concentration achieving 50% lethality or tumour regression). The GI 50 values, a measure of compound's potency, of 2b and 2e along with imatinib 30 are presented in Table 4. Both compounds 2b and 2e showed high potency with one-digit micromolar GI 50 values against all examined cell lines. Interestingly, 2b and 2e elicited superior anticancer potencies than imatinib over 55 and 54 cell lines, respectively. Regarding leukaemia, both 2b and 2e outper-  From another perspective, it was documented that imatinib could ameliorate the hypoxic circumstances in non-small cell lung cancer (NSCLC) 31 , beside its ability to improve drug delivery and efficacy in NSCLC xenografts 32 . In this aspect, it is interesting to observe that both compounds 2b and 2e possess superior potency to imatinib over seven of the tested NSCL cancer cells. For instance, compounds 2b and 2e displayed GI 50 values of 1.47 mM and 1.19 mM against HOP-92 cell line, respectively, with 9-11 folds better potency than imatinib (GI 50 ¼ 13.34 mM). The ethylpiperazine member 2b slightly surpassed the potency of its 3,5bis-trifluoromethyl congener 2e over six cell lines of NSCL cancer panel. Compound 2b exerted equipotent anticancer effects (GI 50 $ 1.76 mM) over NCI-H460 and NCI-H522 cell lines.
On the other hand, imatinib has shown beneficial clinical effects for the treatment of certain metastatic/inoperable gastrointestinal stromal tumours (GIST) 33 . In this aspect, it is noteworthy reporting that both 2b and 2e elicited considerable higher anticancer activities than imatinib over all tested colon cancer cell lines. The growth of HT29, the most imatinib-responsive colorectal cancer cell, was strongly inhibited by 2b and 2e, with GI 50 values of 1.31 mM and 2.15 mM, respectively. While imatinib exerted modest anticancer activity over COLO205, HCC-2998, HCT-116 and multidrug-resistant HCT-15 colon cancer cells (GI 50 ¼ 12.59-21.23 mM), 2b and 2e showed single-digit micromolar GI 50 values spanning in the range of 1.07-2.65 mM. Such promising cellular findings recommend further investigations of 2b and 2e as potential candidates for colorectal cancer.
Since glioblastoma (GBM), the most aggressive form of brain tumours 34 , is resistant to the majority of tyrosine kinase inhibitors, including imatinib 35 , it is important to point out that both 2b and 2e were remarkably superior to imatinib across all six examined brain cancer cell lines. Moreover, 2b and 2e exerted good anticancer effects against the temozolomide (TMZ)-resistant cell lines SF-295, SNB-75, SNB-19, and SF-268 gliomas. For instance, the ethyl piperazine member 2b suppressed the growth of the SNB-19 cell line with a GI 50 value of 2.03 mM. Referring to melanoma, both 2b and 2e showed comparable potencies against all tested melanoma cell lines, surpassing imatinib, with GI 50 values of 1.54-2.51 mM.
Moreover, imatinib efficacy in the treatment of ovarian cancers as a combination with docetaxel was reported 36 . In this line, it is interesting to find that compounds 2b and 2e have significant anticancer activity outperforming imatinib over all tested ovarian cancer cells, except OVCAR-5. Particularly, 2b showed GI 50 values less than 2.0 mM over all cell lines. Of special interest, ureidobenzothiazoles 2b and 2e exerted appreciable potency towards the paclitaxel-resistant NCI/ADR-RES cell line, with GI 50 values of 1.72 mM and 2.37 mM, respectively.
Beside the aforesaid-targeted cellular potencies of 2b and 2e, they elicited broad-spectrum activities over multiple cell lines originating from different cancer tissues. As an example, multidrugresistant (MDR) UO-31 and RXF-393 renal cells showed high sensitivity to 2b (GI 50 ¼ 1.52 mM and 1.70 mM) and 2e (GI 50 ¼ 1.79 mM and 2.19 mM), respectively. Moreover, the top active member 2b  Regarding the efficacy parameters (TGI and LC 50 values) of ureidobenzothiazoles 2b and 2e (Table 5), it was found that compound 2b with hydrophilic ethylpiperazine motif possesses better efficacy than 2e against several cancer cells with TGI < 5.0 mM and LC 50 < 10.0 mM. For example, 2b induced total growth inhibition against K-562 and HCT-15 cell lines, with TGI values of 1.34 and 2.47 mM, respectively.

Molecular docking study
To get insights about the binding mode of this new set of ureidobenzothiazoles, molecular docking for the most potent candidate 2b (AK-HW-90) into the catalytic kinase domain of wild-type BCR-ABL (PDB code: 2GQG) 37 and its T315I mutant (PDB code: 2Z60) 38 , in their DFG-in conformation, was conducted. As shown in Figure  6, compound 2b showed tight binding within the ATP binding site of both BCR-ABL WT and BCR-ABL T315I kinases, as demonstrated by the formation of crucial three hydrogen bonds (HB) with hinge region residue Met318. The ring nitrogen of the benzothiazole scaffold was engaged in HB with amide hydrogen of Met318, while the two NH hydrogens of urea moiety formed two HBs with the amino carbonyl oxygen of Met318. In addition, attractive charge interaction was noticed between the terminal nitrogen of ethyl piperazine and the side chain carboxyl group of Glu329 at the solvent-accessible channel of the BCR-ABL kinase. Moreover, the methoxy group shared carbon-hydrogen bonds with the side chain hydroxyl group of Thr315 in BCR-ABL WT . However, in the case of BCR-ABL T315I , due to steric hindrance with Ile315, which has a larger residual volume than Thr315, such carbon-hydrogen interactions were lacking. Instead, p-alkyl interactions were observed between the methoxy group and the side chain phenyl group of the Tyr253 residue ( Figure S1). Besides, in the case of BCR-ABL WT , p-sulphur interaction was noticed between the sulphur atom of benzothiazole and the phenyl group of the Tyr253 residue. The aforementioned binding interactions might justify the sub-nanomolar potency of 2b against both BCR-ABL WT and its T315 mutant.

In silico bioavailability prediction
The bioavailability of compound 2b was predicted by SwissADME online platform 39 . The bioavailability radar panel evaluates the bioactive candidates in terms of their physicochemical parameters, such as lipophilicity, flexibility, polarity insolubility, size, and saturation. Interestingly, compound 2b was found to be in the pink region (optimal ranges) with favourable drug-like properties ( Figure S2).

Conclusion
In the current investigation, we report the design and synthesis of new phenyl-ureidobenzothiazoles 2a-f in an effort to improve the BCR-ABL inhibitory activity of our previously reported lead AKE-5l. Two major structural modifications of AKE-5l were undertaken herein, replacing the 6-oxypicolinamide moiety and the propyl spacer of AKE-5l with o-methoxyphenyl and phenyl ring, respectively. Among the target compounds, both 2a and its ethylpiperazine congener AK-HW-90 (2b) stood as the most potent members  with IC 50 values of 2.0 and 0.65 nM against BCR-ABL T315I , respectively, surpassing the activity of lead AKE-5l (IC 50 ¼ 39.9 nM). Furthermore, AK-HW-90 (2b) potently inhibited the activity of a panel of imatinib-resistant mutants with single digit nanomolar IC 50 values. On the cellular level, in contrast to the narrow spectrum anticancer activity of the lead AKE-5l, compound AK-HW-90 (2b) exerted considerable broad-spectrum anticancer activity across multiple cell lines including the CML K-562 cell. In addition, AK-HW-90 (2b) surpassed the activity of imatinib over more than 50 cancer cell lines with single digit micromolar GI 50 values. Molecular docking study of AK-HW-90 (2b) was conducted to get insights about its postulated binding mode with BCR-ABL WT and BCR-ABL T315I . Considering the cell-free and cell-based assay results, compound AK-HW-90 (2b) may serve as a promising chemotherapeutic candidate for CML and other cancer types.

Disclosure statement
No potential conflict of interest was reported by the author(s).