Development of novel anilinoquinazoline-based carboxylic acids as non-classical carbonic anhydrase IX and XII inhibitors

Abstract As part of our ongoing endeavour to identify novel inhibitors of cancer-associated CA isoforms IX and XII as possible anticancer candidates, here we describe the design and synthesis of small library of 2-aryl-quinazolin-4-yl aminobenzoic acid derivatives (6a–c, 7a–c, and 8a–c) as new non-classical CA inhibitors. On account of its significance in the anticancer drug discovery and in the development of effective CAIs, the 4-anilinoquinazoline privileged scaffold was exploited in this study. Thereafter, the free carboxylic acid functionality was appended in the ortho (6a–c), meta (7a–c), or para-positon (8a–c) of the anilino motif to furnish the target inhibitors. All compounds were assessed for their inhibitory activities against the hCA I, II (cytosolic), IX, and XII (trans-membrane, tumour-associated) isoforms. Moreover, six quinazolines (6a–c, 7b, and 8a–b) were chosen by the NCI-USA for in vitro anti-proliferative activity evaluation against 59 human cancer cell lines representing nine tumour subpanels.


Introduction
Carbonic anhydrases (CA, EC 4.2.1.1) are ubiquitous metalloenzymes that play a crucial role in catalysing the reversible hydration reaction of carbon dioxide to bicarbonate and protons. 1 This reaction, catalysed by Zn þ2 ion, has a critical role in many physiological and pathological processes such as gluconeogenesis and tumorigenicity. 2,3 So far, fifteen human CA (hCA) isoforms have been identified, with varying distributions across tissues and cells. 4 As a result of the dysfunction of different hCA isoforms activities, a number of pathological repercussions might occur, featuring these hCA isoforms as interesting pharmacological targets for a variety of therapeutic approaches using small molecule CA inhibitors (CAIs). 4 Thus, the pharmacological applications of CAIs are identified for the management of diverse disorders such as ophthalmologic problems, 5 epilepsy, 6 obesity 7 and human malignancies. 8 Sulphonamides and their sulfamides and sulfamate bioisosteres are considered as classical hCA inhibitors with a high affinity to the zinc ion in the active site. 3 It is worth to mention that although identification of several chemotypes of CAIs, like coumarins, phenols, thiocarbamates, and carboxylates, [9][10][11] only primary sulfonamide-tethered CAIs have been clinically used for glaucoma (such as acetazolamide and dorzolamide), and investigated in the clinical trials for the treatment of human malignancies (SLC-0111), Figure 1. 12,13 These sulfonamide-tethered CAIs produce strong CA inhibition, however, a number of them lack the necessary isoform selectivity. So, the design and synthesis of new nonclassical CAIs stands out as a promising strategy to discover effective and isoform-selective CAIs for the management of different diseases. The carboxylic acid-based derivatives represent an important non-classical CAIs chemotype that can exert the CA inhibitory effect through different modes of action, such as anchoring to the zinc-bound water-hydroxide ion through Hbonding, or direct binding to the catalytic zinc displacing bound water-hydroxide anion. [14][15][16] In the few last years, we have reported several carboxylic acidtethered small molecules as new CAIs. [17][18][19] A novel series of benzofuran-based carboxylic acids was described as promising CA inhibitors in 2020. 20 Among these benzofuran derivatives, compound I (Figure 1) with a meta-benzoic acid moiety inhibited hCA IX at a submicromolar concentration (K I ¼ 0.79 lM), as well as exerted good hCA XII inhibitory activity (K I ¼ 2.3 lM). Also in the same year, we have developed a small library of methylthiazolo[3,2-a]benzimidazole-based carboxylic acid derivatives as novel CA inhibitors. 21 In particular, compound II ( Figure 1) effectively suppressed CA isoforms IX and XII with inhibition constants equal 0.83 lM and 2.4 lM, respectively. Furthermore, we identified a new series of non-classical CA inhibitors that incorporates enaminone-based carboxylic acids. 22 Compound III (Figure 1) endowed with a para-benzoic acid motif showed submicromolar hCA IX inhibitory activity (K I ¼ 0.92 mM) and good hCA XII inhibitory activity (K I ¼ 1.1 mM).
Based on the findings described above, and as part of our ongoing endeavour to identify novel inhibitors of cancer-associated CA isoforms IX and XII as possible anticancer candidates, [23][24][25][26][27][28][29] here we describe the design and synthesis of a small library of 2aryl-quinazolin-4-yl aminobenzoic acid derivatives (6a-c, 7a-c, and 8a-c) as new non-classical CA inhibitors ( Figure 1). On account of its significance in the anticancer drug discovery and development, [30][31][32][33] and in the development of effective CAIs, 34,35 the 4-anilinoquinazoline privileged scaffold was exploited in this study. Thereafter, the free carboxylic acid functionality was appended in the ortho (6a-c), meta (7a-c), or para-positon (8a-c) of the anilino motif to furnish the target inhibitors.
All the newly synthesised quinazoline-based carboxylic acid derivatives (6a-c, 7a-c, and 8a-c) were assessed for their inhibitory activities against the hCA I, II (cytosolic), IX and XII (trans membrane, tumour associated) isoforms by the stopped-flow CO 2 hydrase assay. Moreover, six quinazolines (6a-c, 7b, and 8a-b) were chosen by the NCI-USA for in vitro anti-proliferative activity evaluation against 59 human cancer cell lines representing nine tumour subpanels.

Chemistry
Melting points ( C, uncorrected) were determined using a Stuart melting point apparatus. The IR spectra (KBr) were recorded on a SHIMADZU FT/IR spectrometer. The NMR spectra recorded by BRUKER 400 MHz NMR spectrometers using DMSO-d 6 as the solvent. Chemical shifts were reported in parts per million (d), and coupling constants (J) expressed in Hertz. 1 H and 13 C spectra were run at 400 and 101 MHz, respectively. Microanalytical data (C, H, and N) were obtained by FLASH 2000 CHNS/O analyser.
General procedures for the synthesis of 2-aryl-4-chloroquinazolines (4a-c) To a suspension of 2-arylquinazolinones 3a-c (1eq) in phosphorus oxychloride (10 eq), a catalytic amount of DMF was added. 36 The reaction mixture was then heated at 90 C for 4h. After cooling, the mixture was added drop-wise to ice-water with stirring, neutralised by ammonium hydroxide, and extracted by methylene chloride. The organic layer was washed with cold water, dried over anhydrous Na 2 SO 4 , and evaporated in vacuo. The obtained solid was crystallised from isopropanol to afford the key 2-aryl-4chloroquinazolines intermediates 4a-c.

Biological evaluation
CA inhibitory assay All the newly synthesised quinazoline-based carboxylic acid derivatives (6a-c, 7a-c and 8a-c) were assessed for their CA catalysed CO 2 hydration activities against hCA isoforms I, II, IX and XII by the stopped flow CO 2 hydrase assay as reported previously [37][38][39][40] (Supporting Materials).
In vitro NCI-59 cancer cell lines assays The NCI-USA anticancer assays was performed utilising the NCI, Bethesda, Drug Evaluation Branch protocol, 41-43 using the SRB cytotoxicity assay, 44 as desribed earlier. 45,46 Results and discussion
The target quinazoline derivatives (6a-c, 7a-c and 8a-c) were structurally confirmed by spectral and elemental analyses. The 1 H NMR spectra of all compounds revealed a singlet signal around d 11.70-12.34 ppm due to the proton of the NH group. Moreover, all compounds showed two doublet signals in the aromatic region around d 7.49-8.40 and 8.61-9.03 ppm that are attributable to H5 and H8 of quinazoline moiety, respectively. In addition, 1 H NMR spectra for derivatives (6a-b, 7a-b and 8a-b) showed another singlet signal for the CH 3 group at the range of d 2.41-2.43 ppm, whereas, the 1 H NMR spectra for (6c, 7c and 8c) disclosed the singlet signal of the OCH 3 group around d 3.87-3.90 ppm. One the other hand, 13 C NMR spectra for the target quinazoline derivatives confirmed the presence of the carboxylic C¼O functionality at d 162-170 ppm. Furthermore, 13 C NMR spectra for compounds (6ab, 7a-b, and 8a-b) showed a signal at d 21.41-21.70 ppm for the CH 3 carbon, whereas spectra of compounds (6c, 7c, and 8c) displayed a signal at d 56.02-56.25 ppm for the OCH 3 carbon.

Biological evaluation
Carbonic anhydrase inhibition All the newly synthesised quinazoline-based carboxylic acid derivatives (6a-c, 7a-c, and 8a-c) were assessed for their inhibitory activities against the hCA I, II (cytosolic), IX and XII (trans membrane, tumour associated) isoforms by the stopped-flow CO 2 hydrase assay. 37 Acetazolamide (AAZ) was used as a standard CA inhibitor. The data is summarised in Table 1.
Only three of the tested quinazoline-based carboxylic acids (8a, 8b, and 8c) weakly inhibited the cytosolic hCA I isoform, with inhibition constants (K I s) equal 87.7, 73.2, and 66.3 mM, respectively, whereas quinazoline derivatives 6a-c and 7a-c could not inhibit hCA I up to 100 mM. These results revealed that grafting the carboxylic acids functionality at the para position (8a-c) could result in modest hCA I inhibitory activity, while shifting to ortho-(6a-c) or meta-(7a-c) positions resulted in the elimination of hCA I inhibitory activity (K I s > 100 M), Table 1. The cytosolic hCA II was effectively inhibited by para-aminobenzoic acid-bearing quinazolines (8a-c) with K I s of 9.3, 3.9 and 4.6 mM, respectively, whereas, their ortho (6a-c) and meta (7a-c) regioisomers elicited modest inhibitory effects with inhibition constants spanning in the range of 26.0 À 85.8 mM. It is worth to mention that substitution of the 2-phenyl motif with a 4-methyl group, in series 8, led to compound 8b with the best hCA II inhibitory activity (K I ¼ 3.9 mM).
Similar to the hCA I and hCA II inhibition profiles, the obtained K I values disclosed that the cancer-related hCA IX isoform was inhibited most effectively by para-aminobenzoic acid-bearing quinazolines (8a-c) with K I s equal 4.3, 1.6 and 4.5 mM, respectively. In addition, hCA IX was moderately affected by quinazolines decorated with ortho and meta aminobenzoic acid motifs with K I s ranging between 24.2 and 46.4 mM. The order of activities of target quinazoline-based carboxylic acids towards hCA IX was increased in the order of para isomers 8 (K I s: 1.6 À 4.5 mM) > meta isomers 7 (K I s: 24.2 À 31.6 mM) > ortho isomers 6 (K I s: 34.4 À 46.4 mM), Table 1. Regarding the impact of substitution on the 2-phenyl moiety, within series 6, 7, and 8, it was found that the order of hCA IX inhibitory activities was 4-methyl derivatives (6b, 7b, and 8b; K I s ¼ 34.4, 24.2 and 1.6 mM) > 3-methyl derivatives (6a, 7a, and 8a; K I s ¼ 42.5, 29.3 and 4.3 mM) > 4-methoxy derivatives (6c, 7c and 8c; K I s ¼ 46.4, 31.6 and 4.5 mM), Table 1.
The second cancer-related isoform studied in this study is hCA XII, which is also the most vulnerable to the prepared molecules. All quinazoline-based carboxylic acids (6a-c, 7a-c, and 8a-c) exhibited good inhibition of hCA XII (K I s: 0.25 À 9.0 mM), as seen by the data in Table 1. In particular, the best hCA XII inhibitory effect was exerted by quinazoline 8c with a K I value equals 0.25 mM. Besides, quinazolines 7b, 7c, and 8b displayed also submicromolar inhibitory activity towards hCA XII with K I values 0.91, 0.48, and 0.42 mM, respectively. Similarly to the abovementioned deduced Structure-Activity Relationship (SAR) for hCA I, II, and IX isoforms, the order of hCA XII inhibitory activities was increased in the order of para isomers 8 (K I s: 0.25 À 3.8 mM) > meta isomers 7 (K I s: 0.48 À 4.8 mM) > ortho isomers 6 (K I s: 7.1-9 mM), Table 1. Also, it's noteworthy that appending p-methoxyphenyl moiety at C-2 of quianzoline within series 7 and 8 (7c and 8c; K I s ¼ 0.48 and 0.25 mM) resulted in a better hCA XII inhibitory activity than pmethylphenyl (7b and 8b; K I s ¼ 0.91 and 0.42 mM) and m-methylphenyl (7a and 8a; K I s ¼ 4.8 and 3.8 mM) moieties. The SAR for the inhibitory activity of the new quinazolines towards different hCA isoforms is summarised in Figure 2.
As a result of the inhibitory profile for the reported quinazoline-based carboxylic acid derivatives (6a-c, 7a-c, and 8a-c) ( Table 1), the selectivity index (SI) for each derivative was calculated and presented in Table 2. Regarding the selectivity towards CA IX and XII isoforms, most the examined quinazoline-based carboxylic acids (6a-c, 7a-c, and 8a-c) exhibited low to moderate selectivity, except compounds 7b and 7c that disclosed excellent selectivity towards hCA XII over hCA I (SI ¼ 109.9 and 208.3, respectively), and hCA II (SI ¼ 45.82 and 178.75, respectively), in addition, compounds 8b and 8c displayed outstanding selectivity towards hCA XII over hCA I (SI ¼ 174.3 and 265.2, respectively), Table 2.

In vitro anti-proliferative activity
The structures of all novel quinazoline based-carboxylic acids prepared in this study were submitted to the National Cancer Institute (NCI-USA), and six compounds (6a-c, 7b, and 8a-b) were chosen for in vitro anti-proliferative activity evaluation against fifty-nine human cancer cell lines representing nine tumour subpanels, according to Bethesda, Drug Evaluation Branch Protocol. [41][42][43] Preliminary single dose screening at 10 mM concentration. The anti-proliferative activities of the selected quinazoline derivatives (6a-c, 7b, and 8a-b) have been evaluated at single (10 lM) dose assay using the SRB protocol. 44 The obtained data was presented as a mean-graph of the percentage growth of the various treated cancer cells and was displayed in Table 2 as the percentage growth inhibition (GI%) induced by the investigated compounds.
Examining the data in Table 3 revealed that the tested quinazoline-based-carboxylic acids (6a-c, 7b, and 8a-b) demonstrated diverse patterns of sensitivity and selectivity against the various NCI cancer cell panels. Quinazoline derivatives featuring ortho aminobenzoic acid (6a-c) showed excellent broad cell growth inhibitory activity (GI % mean ¼ 63, 84, and 52, respectively) against most of all cancer cell lines, whereas compounds (7b and 8a-b) with meta and para-aminobenzoic acid moiety showed fair selective activity (GI % mean ¼ 14, 16 and 20, respectively) towards certain cancer cell lines as shown in Table 3.
In vitro NCI 5-dose assay. The preliminary screening data showed that quinazoline-carboxylic acid 6b (NSC: 835857) was the most active anticancer molecule in this study, with promising activity against numerous tumour cell lines. Thus, 6b was selected by NCI for further evaluations at a 5-doses (0.01-100 mM) level. Three dose-response parameters (GI 50 , TGI, and LC 50 ) were calculated and displayed in Table 4.
Concerning  (Table 3). Table 2. Selectivity ratios for the inhibition of CA IX and XII isoforms over CA I and II isoforms for carboxylic acids (6a-c, 7a-c, and 8a-c) and acetazolamide.  Table 3. Percentage growth inhibition (GI%) of subpanel tumour cell lines at 10 lM concentration of the quinazoline based-carboxylic acids (6a-c, 7b, and 8a-b). With regard to the sensitivity of the examined cell lines, quinazoline 6b elicited comparatively homogenous growth inhibitory activity throughout all NCI panels, with good growth inhibition full panel GI 50 (MG-MID) equals 11.99 lM, as well as subpanel GI 50 (MG-MID) values spanning from 8.04 to 15.66 lM. In particular, the most susceptible subpanels were CNS and Leukaemia with MG-MID of 8.04 and 8.68 lM, respectively (Table 5). In order to assess the selectivity of 6b, its full panel MG-MID is divided by its individual subpanel MG-MID ( Table 5). The selectivity index for compound 6b ranged from 0.76 to 1.49 which points out that 6b has non-selective broad-spectrum anti-proliferative activity towards all NCI cancer subpanels. It is worth to mention that the best antiproliferative counterpart 6b is not the most active inhibitor against CA IX or XII, thus the target of this compound could be other than CAs.

Conclusions
Three sets of 2-aryl-quinazolin-4-yl aminobenzoic acid regioisomers (6a-c, 7a-c, and 8a-c) were designed and synthesised as new non-classical CA inhibitors. Their CA inhibitory activities towards isoforms I, II, IX, and XII were evaluated. Only three of the tested quinazoline-based carboxylic acids (8a, 8b, and 8c) weakly inhibited the cytosolic hCA I isoform, with inhibition constants (K I s) equal 87.7, 73.2, and 66.3 mM. The cytosolic hCA II was effectively inhibited by para-aminobenzoic acid-bearing quinazolines (8a-c) with K I s of 9.3, 3.9, and 4.6 mM, respectively, whereas, their ortho (6a-c) and meta (7a-c) regioisomers elicited modest inhibitory effects. Moreover, the cancer-related hCA IX isoform was inhibited most effectively by quinazolines (8a-c) with K I s equal 4.3, 1.6, and 4.5 mM, respectively. Also, the results revealed that the cancerrelated hCA XII isoform is the most vulnerable to the prepared molecules. In particular, the best hCA XII inhibitory effect was exerted by quinazoline 8c (K I ¼ 0.25 mM), also, quinazolines 7b, 7c, and and 8b displayed sub-micromolar hCA XII inhibitory activity (K I ¼ 0.91, 0.48, and 0.42 mM, respectively). The SAR analysis highlighted that the order of hCA inhibitory activities was increased in the order of para isomers 8 > meta isomers 7 > ortho isomers 6. On the other hand, the anti-proliferative activities of the quinazoline derivatives (6a-c, 7b, and 8a-b) have been evaluated at single (10 lM) dose assay against 59 cancer cell lines in the NCI-USA. Quinazoline derivatives featuring ortho aminobenzoic acid (6a-c) showed excellent broad cell growth inhibitory activity (GI % mean ¼ 63, 84 and 52, respectively) against most of all cancer cell lines, whereas compounds (7b and 8a-b) with meta and para aminobenzoic acid moiety showed fair selective activity (GI

Disclosure statement
CT Supuran is Editor-in-Chief of the Journal of Enzyme Inhibition and Medicinal Chemistry. He was not involved in the assessment, peer review, or decision-making process of this paper. The authors have no relevant affiliations of financial involvement with any organisation or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Funding
The author(s) reported there is no funding associated with the work featured in this article.