Comparative Analysis of Structural Analogs of Dipyridothiazines with m-Xylene and a Lutidine Moiety—In Silico, In Vitro, and Docking Studies

Abstract

Dimers of dipyridothiazines with an m-xylene moiety are presented in terms of a comparative analysis with anticancer active structural analogs containing a lutidine system. The synthesis of new isomeric dimers was described, the structure of which was confirmed by ¹H, ¹³C and 2D NMR, and HR MS spectroscopic methods. The preliminary prediction of the pharmacological profile using the Way2Drug server indicated the anticancer potential of the tested derivatives. In vitro biological activity tests were performed on a normal skin cell line (HaCaT) and five cancer cell lines, including human primary colon cancer (SW480), human metastatic colon cancer (SW620), human breast adenocarcinoma (MDA-MB-231), human lung carcinoma (A-549), and human glioblastoma (LN-229), which indicated low cytotoxic activity. In order to explain the surprisingly low activity, a comparative structural analysis of the tested analogs compared to the dimers with the lutidine system was performed using quantum mechanics and molecular docking in relation to histone deacetylase. Molecular docking indicated the different binding sites of the analyzed dimers, which explained the differences in the activity.

1. Introduction

Phenothiazines are completely synthetic heterocyclic systems that have no precursor or equivalent in the world of living nature [1,2]. Although they were discovered more than 140 years ago, to this day, the molecules are still very popular in the world of medical, pharmaceutical, as well as chemical research [3,4,5].

In the 1950s, phenothiazines containing dialkylaminoalkyl substituents in their structure triggered revolutionary changes in psychiatry by showing phenomenal antipsychotic activity to treat schizophrenia or manic states. To this day, these drugs are still used in medicine [1]. Scientific research in the area of phenothiazines is constantly being carried out, revealing many further interesting pharmacological activities, including anticancer [6,7,8,9,10], antimicrobial [11,12], antifungal [13], antiviral [14,15], antioxidant [16], or the reversal of multidrug resistance activities [17]. This topic is so rich that its results are published in numerous review papers each year around the world [18,19,20,21,22]. In 2006, English researcher Mitchell named the phenothiazine “parent molecule”, which fully reflects the application nature of this chemical compound [22].

The structure of phenothiazines is also modified at various levels, which leads to the creation of new, previously undescribed derivatives with interesting biological activities [23]. Such a group of modified phenothiazines includes dipyridothiazines having two pyridine rings in their structure instead of benzene rings [21]. Dipyridothiazines have a valuable anticancer effect against numerous cancers, such as melanoma, glioma, lung, colon, breast, and ovarian cancer, consisting of the induction of apoptosis through its mitochondrial activation and the possibility of partial DNA intercalation [21,24]. Dipyridothiazines also showed significant immunosuppressive and anti-inflammatory activity by inhibiting TNF alpha and interleukin 6 levels, as well as low cytotoxicity towards normal cells [21].

Dimeric derivatives of dipyridothiazines, in which two dipyridothiazine units were connected by selected organic linkers, such as o-xylene, p-xylene, and 2,6-dimethylpyridine (lutidine), have recently shown highly promising anticancer activities in relation to colon (SW480) and breast (MCF 7) cancer lines. Additionally, these compounds were characterized by their relatively low cytotoxicity in relation to normal myoblast (L6) cell lines [25]. Among dimeric dipyridothiazine derivatives, the highest anticancer potential was demonstrated by compounds that contained a linker in their structure, which was 2,6-dimethylpyridine, i.e., lutidine. Additionally, pyridine and its substituted derivatives are known as six-membered heterocycles with broad pharmacological activity profiles [26,27].

The idea of this work was to synthesize new dimeric derivatives of dipyridothiazines (1,6-, 1,8-, 2,7-, and 3,6-diazaphenothiazines 1–4) with an m-xylene linker (6–9) as structural analogs of the early described dimers with the lutidine moiety (6a–9a) [25] (Scheme 1, Figure 1) in order to comparatively analyze their biological activity and determine the structure–activity relationship in the area of in vitro anticancer research.

The structure of the new compounds was clearly confirmed by the NMR, 2D NMR spectroscopy, and mass spectrometry HR MS. In silico analyses of probable molecular targets were performed using the Way2Drug server. The compounds were tested for anticancer activity in relation to the following cancer cell lines: human primary colon cancer (SW480), human metastatic colon cancer (SW620), human breast adenocarcinoma (MDA-MB-231), human lung carcinoma (A-549), and human glioblastoma (LN-229), and the human immortal keratinocyte cell line from adult human skin (HaCaT). To elucidate the structure–activity relationship, quantum mechanical calculations and a molecular docking study were performed against the selected histone deacetylase target.

2. Materials and Methods

2.1. Chemistry

The NMR spectra were recorded on Bruker Advance spectrometers (¹H at 600 MHz, ¹³C at 150 MHz) in CDCl₃ and DMSO_d6. Two-dimensional COSY, ROESY, HSQC, and HMBC spectra of compound 8 were recorded on a Bruker Advance spectrometer at 600 MHz. All NMR analyses were performed at room temperature using the standard TopSpin 3.7.0 software. The HR MS spectra (ESI—Electro Spray Ionisation) were run on a Brucker Impact II. The samples were prepared as solutions in acetonitrile. Melting points were determined in open capillary tubes on a Boetius melting point apparatus. TLC was performed on silica gel 60 F₂₅₄ (Merck 1.05735) with CHCl₃-EtOH (10:1 v/v) and on Al₂O₃ 60 F₂₅₄ neutral (type E) (Merck 1.05581) with CHCl₃-EtOH (10:1 v/v) as eluents.

The parent starting dipyridothiazines 1–4 (Scheme 1) and lutidine derivatives 6a–9a were prepared according to reference [25].

General Procedure for Synthesis of Compounds (6–9)

To a solution of the selected diazaphenothiazine (1–4) (0.201 mg, 1 mmol) in dry DMF (10 mL), NaH (0.028 g, 1.2 mmol, 60% NaH in mineral oil was washed out with hexane) was added. The reaction mixture was stirred at room temperature for 1 h and then the linker α,α′-dichloro-m-xylene (0.5 mmol) was added, and the stirring was continued for 24 h. The progress of the reaction was monitored using TLC. The mixture was poured into water (15 mL), extracted with CHCl₃ (3 × 10 mL), and dried using Na₂SO₄. The obtained product was purified by column chromatography (aluminum oxide, CHCl₃-Ethanol 10:1) to give the following:

1,3-bis((10H-dipyrido [2,3-b:2′,3′-e][1,4]thiazin-10-yl)methyl)benzene (6)

(214 mg, 85%), mp = 268–270 °C

¹H NMR (DMSO_d6) δ: 5.16 (s, 4H, 2 CH₂), 6.69 (dd, J = 8.4 Hz, J = 1.2 Hz, 2H), 6.78 (dd, J = 7.2 Hz, J = 4.8 Hz, 2H), 6.84 (dd, J = 8.4 Hz, J = 4.8 Hz, 2H), 6.93 (s, 1H, H_d-m-xylen), 7.18 (d, J = 7.2 Hz, 2H, H_b-m-xylen), 7.31 (t, J = 7.2 Hz, 1H, H_c-m-xylen), 7.37 (dd, J = 7.2 Hz, J = 1.2 Hz, 2H), 7.68 (dd, J = 4.8 Hz, J = 1.2 Hz, 2H), and 7.82 (dd, J = 4.8 Hz, J = 1.2 Hz, 2H). ¹³C NMR: 47.58, 115.14, 119.10, 121.95, 122.56, 124.59, 125.45, 129.28, 135.11, 137.16, 137.92, 142.85, 143.17, 145.41, and 151.89. HR MS (EI) m/z for [C₂₈H₂₁N₆S₂ + H], calc. 505.1264; found: 505.1283 (100%).

1,3-bis((10H-dipyrido [3,2-b:3′,4′-e][1,4]thiazin-10-yl)methyl)benzene (7)

(209 mg, 83%), mp = 262–264 °C

¹H NMR (CDCl₃) δ: 5.42 (s, 4H, 2 CH₂), 6.90 (m, 2H), 7.17 (d, J = 6.0 Hz, 2H), 7.18 (d, J = 7.8 Hz, 2 H_b-m-xylen), 7.21 (m, 2H), 7.27 (m, 1H, H_c-m-xylen), 7.70 (s, 1H, H_d-m-xylen), 8.05 (m, 2H), 8.14 (d, J = 6.0 Hz, 2H), and 8.19 (s, 2H). ¹³C NMR: 48.02, 111.95, 120.20, 121.59, 125.52, 126.67, 128.23, 130.03, 134.95, 135.05, 140.91, 143.84, 147.14, 151.13, and 152.90. HR MS (EI) m/z for [C₂₈H₂₁N₆S₂ + H], calc. 505.1264; found: 505.1283 (100%).

1,3-bis((10H-dipyrido [3,4-b:3′,4′-e][1,4]thiazin-10-yl)methyl)benzene (8)

(216 mg, 86%), mp = 256–257 °C

¹H NMR (CDCl₃) δ: 5.06 (s, 4H, 2 CH₂), 6.35 (d, J = 5.7 Hz, 2H, 2 H₉), 6.94 (s, 1H, H_d-m-xylen), 6.97 (d, J = 4.8 Hz, 2H, 2 H₄), 7.25 (d, J = 7.2 Hz, 2H, 2 H_b-m-xylen), 7.45 (t, J = 7.2 Hz, 1H, H_c-m-xylen), 7.70 (s, 2H, 2 H₁), 7.96 (d, J = 5.7 Hz, 2H, 2 H₈), 8.01 (s, 2H, 2 H₆), and 8.10 (d, J = 4.8 Hz, 2H, 2 H₃). ¹³C NMR: 51.08 C_CH2, 109.41 C₉, 117.19 C_9a, 121.56 C_d, 123.48 C₄, 126. 06 C_b, 130.79 C_c, 133.35 C_10a, 134.46 C_a, 135.56 C₁, 137.22 C_4a, 143.44 C₆, 145.70 C₃, 147.14 C₈, and 151.47 C_5a. HR MS (EI) m/z for [C₂₈H₂₁N₆S₂ + H], calc. 505.1264; found: 505.1229 (100%).

1,3-bis((10H-dipyrido [2,3-b:4′,3′-e][1,4]thiazin-5-yl)methyl)benzene (9)

(199 mg, 79%), mp = 288–290 °C

¹H NMR (DMSO_d6) δ: 5.04 (s, 4H, 2 CH₂), 6.44 (d, J = 5.4 Hz, 2H), 6.81 (m, 2H), 6.88 (m, 2H), 6.97 (s, 1H, H_d-m-xylen), 7.27 (d, J = 7.2 Hz, 2H, H_b-m-xylen), 7.44 (t, J = 7.2 Hz, 1H, H_c-m-xylen), 7.90 (d, J = 4.2 Hz, 2H), 7.95 (d, J = 5.4 Hz, 2H), and 8.02 (s, 2H). ¹³C NMR (CDCl₃): 50.10, 110.06, 117.65, 121.69, 128.61, 129.01, 131.28, 134.04, 136.33, 137.99, 145.02, 146.85, 149.14, and 149.63. HR MS (EI) m/z for [C₂₈H₂₁N₆S₂ + H], calc. 505.1264; found: 505.1299 (100%).

2.2. In Silico Target Prediction

The prediction of the biological targets and cytotoxicity towards the cancer cell lines was performed using the Way2Drug server [28,29]. This server analyzes the biological activity of an organic drug based on molecular recognition against its database.

2.3. Biological Evaluation

2.3.1. Cell Line and Culture

Human primary colon cancer (SW480) and metastatic colon cancer (SW620), human breast adenocarcinoma (MDA-MB-231), human lung carcinoma (A-549), human glioblastoma (LN-229), and human immortal keratinocyte (HaCaT) cell lines were obtained from the American Type Culture Collection (ATCC) in Rockville, MD, USA. The following cell lines were cultured in the recommended medium according to ATCC protocols: SW480 and SW620 in Minimum Essential Medium (MEM), and MDA-MB-231, A-549, LN-229, and HaCaT cells in Dulbecco’s Modified Eagle Medium (DMEM). The culture medium was supplemented with 10% fetal bovine serum (FBS), penicillin (100 U/mL), and streptomycin (100 μg/mL). Cells were maintained in a humidified incubator at 37 °C with 5% CO₂. Passaging was performed when the cells reached 80–90% confluence using 0.25% trypsin (Gibco Life Technologies, Gibco, Grand, Island, NY, USA) treatment before being utilized for the experiments.

2.3.2. MTT Cell Viability Assay

Cell viability was determined through the enzymatic conversion of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide salt (MTT) to insoluble formazan crystals by the mitochondrial dehydrogenases present in the viable cells [30]. Cells were seeded into 96-well plates at a density of 1 × 10⁴ cells per well and allowed to adhere for 24 h at 37 °C in a CO₂ incubator. Following the 24 h incubation period, the culture medium was replaced with fresh medium, and the cells were exposed to varying concentrations of the compounds (ranging from 10 to 100 μM). Then, they were further incubated for 72 h at 37 °C in a CO₂ incubator. Untreated cells served as controls. Subsequently, MTT solution (0.5 mg/mL in serum-free medium) was added, and the samples were incubated for 4 h at 37 °C in a CO₂ incubator. Following this, the medium was removed, and formazan crystals were solubilized by adding an isopropanol-DMSO mixture (1:1). The absorbance of the dissolved crystals was measured at a wavelength of 570 nm using a UVM 340 reader (ASYS Hitech GmbH, Eugendorf, Austria). The IC₅₀ values were calculated using GraphPad Prism 8 software (GraphPad Software).

2.4. The Statistical Analysis

The statistical analysis was performed using GraphPad Prism 9 [31] software (Graph Pad Software, San Diego, CA, USA). The results were reported as mean ± SD from a minimum of three independent experiments. Statistical significance of the differences between the values was assessed using an analysis of variance with Dunnett’s multiple comparison post hoc test, and significance was defined as p < 0.05.

2.5. Quantum Mechanical Calculations

Electron density and bond length calculations were performed using the standard B3LYP functional set and the 6-31+G(d,p) basis set, using the Gaussian16 program [32], and visualized using the widely used Avogadro program [33].

2.6. Molecular Docking Analysis

• Protein preparation

The HDAC4 histone deacetylase structure (PDB ID: 2VQJ) was used for docking. The HDAC4 structure was crystallized with the inhibitor, which was manually removed from the structure. The protonation of the enzyme was performed with an H++ server at a pH of 7.5 [34,35].

• Ligand preparation

The structures of two ligands, 8 and 8a, were created using MarvinSchetch 19.22.0 software. File conversion, including Gasteiger charge assignment, was performed using the OpenBabel 2.3.2 tool [36].

• Molecular Docking

The first docking was carried out using the DiffDock software Version 1.1.2 [37]. Blind docking on default settings was used. Docking was also performed using AutoDock Vina software Version 4.2.6 [38,39] implemented on the TAMARIND BIO server. The docking site was defined as the inhibitor binding pocket in the crystallographic structure. This pocket was also detected during the blind docking using DiffDock software Version 1.1.2, during the analysis of the selected ligands.

3. Results and Discussion

3.1. Chemistry Part

The preparation of new dipyridothiazine dimers was carried out using previously synthesized, selected dipyridothiazines (10H-1,6-, 1,8-, 2,7-, and 3,6-diazaphenothiazines 1–4), for which the promising anticancer activities were determined [25]. In the next step, selected dipyridothiazines were subjected to alkylation reactions with α,α′-dichloro-m-xylene. The reactions were carried out in anhydrous DMF in the presence of the strong base NaH at room temperature, which led to the formation of the final dimers 6–9 (Scheme 1). The obtained dimers were the structural analogs of the dimers with the lutidine moiety, which do not have the nitrogen atom characteristic of the lutidine system. The progress of the reaction was monitored using thin-layer chromatography (TLC). The crude solid reaction products were purified using column chromatography. All of the new compounds were obtained in good yields (79–86%).

3.2. Structural Identification Using Spectroscopic Methods

The identification of the structure of new organic compounds is a fundamental problem both in organic chemistry and in the drug design process [40,41]. Due to the fact that chemical reactions may be accompanied by rearrangements or parallel or subsequent reactions, proving the structure of the obtained product is essential. For this purpose, advanced NMR spectroscopic techniques and high-resolution HR MS mass spectrometry were used. The identification of the molecule’s structure was carried out on the basis of the following ¹H, ¹³C NMR spectra, and 2D NMR correlations: COSY (COrrelation SpectroscopY), ROESY (Rotating frame nuclear Overhauser Effect SpectroscopY), HSQC (Heteronuclear Single Quantum Correlation), HMBC (Heteronuclear Multiple Bond Correlation), and HR MS mass spectrometry. The full structural analysis performed for dimer 8 allowed for the assignment of individual hydrogen and carbon atoms in the molecules. The NMR spectra confirmed the correct number of both protons and carbon atoms in the new compounds. The molecular mass and purity of the compounds were confirmed by high-resolution mass spectrometry (HR MS). All of the spectra of compounds 6–9 are included in Supplementary Materials File S1.

3.3. In Silico Target Prediction

The new dimers 6–9 were analyzed in silico using the Way2Drug server, which includes the PASS (Prediction of Activity Spectra for Substances) application, which allows for the prediction of the potential biological activities of the molecules and the probability of cytotoxicity toward cancer cell lines [28,29]. The obtained results are summarized in

Table 1. Probability (%) of activity spectra of dimers (6–9) using the PASS program.

No.	Probability of Activity Spectrum
6	(28%) Histone deacetylase stimulant	(77%) Glycosylphosphatidylinositol phospholipase D inhibitor	(47%) Transcription factor inhibitor	(33%) Alzheimer’s disease treatment	(46%) Antiallergic
7	(29%) Mitochondrial processing peptidase inhibitor	(83%) Glycosylphosphatidylinositol phospholipase D inhibitor	(18%) Antipsychotic	(30%) Cytochrome P450 inhibitor	(10%) MAP3K8 inhibitor
8	(75%) Neurodegenerative diseases treatment	(84%) Glycosylphosphatidylinositol phospholipase D inhibitor	(65%) Histone deacetylase stimulant	(42%) Alzheimer’s disease treatment	(41%) Cytochrome P450 inhibitor
9	(76%) Histone deacetylase stimulant	(66%) Antiallergic	(60%) Neurodegenerative diseases treatment	(58%) Alzheimer’s disease treatment	(59%) Antiasthmatic

and

Table 2. The probable cytotoxic effect (%) on various cancer cell lines of dimers (6–9).

No.	Probability of Cytotoxicity towards Cancer Cell Lines
6	Breast cancer MDA-MB-468 (62%)	Melanoma UACC-257 (51%)	Non-small-cell lung carcinoma NCI-H322M (49%)	Colon carcinoma HCT-116 (42%)	Pancreatic carcinoma YAPC (43%)
7	Breast cancer MDA-MB-468 (59%)	Melanoma UACC-257 (55%)	Renal carcinoma 786 (45%)	Colon carcinoma HCT-116 (39%)	Pancreatic carcinoma YAPC (44%)
8	Breast cancer MDA-MB-468 (39%)	Melanoma UACC-257 (54%)	Renal carcinoma 786-0 (49%)	Colon carcinoma HCT-116 (55%)	Ovarian adenocarcinoma OVCAR-4 (43%)
9	Breast cancer MDA-MB-468 (44%)	Melanoma UACC-257 (49%)	Renal carcinoma 786-0 (44%)	Colon carcinoma HCT-116 (54%)	Ovarian adenocarcinoma OVCAR-4 (37%)

.

The results show that the new dimers 6–9 have a high probability of biological action aimed at the anticancer potential related to the mechanism of histone deacetylase stimulation, the inhibition of glycosylphosphatidylinositol phospholipase D, which may directly lead to the activation of the mitochondrial apoptosis pathway. It is worth noting that the anticancer potentials of dipyridothiazines 1–4 were also experimentally confirmed in our earlier studies, which was the beginning of the search for new derivatives in the group of dipyridothiazines with even better activities [21,24]. Moreover, previous in silico analyses performed for dimers with the lutidine, o-xylene, and p-xylene systems also indicated a high potential for anticancer activity, which was confirmed [25].

Furthermore, the PASS program indicated the likelihood of dimer activity in neurodegenerative diseases, antipsychotic and antiallergic effects, and the inhibition of cytochrome 450.

More importantly, in addition to the probable mechanisms of biological action, the PASS program indicated, for the new dimeric derivatives, the likelihood of cytotoxic effects on breast, kidney, colon, pancreatic, and ovarian cancer cell lines, melanoma, and non-small-cell lung cancer. It is worth pointing out that previous studies on dipyridothiazines 1–4 confirmed this promising anticancer potential [21], which was an inspiration to conduct further cytotoxicity tests of the new dimers 6–9.

3.4. Biological Part—Analysis of Cytotoxicity towards Cancer and Normal Cells

Taking into account the results of the in silico simulations and the previously promising anticancer activities and low cytotoxicity of the obtained dipirydothiazines 1–4 [21] and dimers with a lutidine linker 6a–9a [25], the new derivatives 6–9 were tested for their anticancer activity. The cytotoxicity and antiproliferative activity of new dimers were determined in vitro in relation to the following five cancer cell lines: human primary colon cancer (SW480), human metastatic colon cancer (SW620), human breast adenocarcinoma (MDA-MB-231), human lung carcinoma (A-549), and human glioblastoma (LN-229), and to the normal line human immortal keratinocyte cell line from adult human skin (HaCaT) using the MTT assay [30]. The study used doxorubicin and cisplatin as reference drugs. The data are expressed as the mean SD, IC₅₀ (µM)—the half-maximal growth inhibitory concentration after culturing the cells for 72 h with the individual compound—and the SI (Selectivity Index), which was calculated using the following formula: SI = IC₅₀ for normal cell line/IC₅₀ cancer cell line. The results are summarized in

Table 3. Cytotoxic activity (IC₅₀, µM) of the studied compounds estimated by the MTT assay.

No.	Cancer Cells										Normal Cells
	SW480		SW620		MDA-MB-231		A-549		LN-229		HaCaT
	IC50	SI	IC50	SI	IC50	SI	IC50	SI	IC50	SI	IC50
6	92.6 ± 5.31	1.08	>100	1	83.2 ± 3.38	1.20	81.5 ± 5.71	1.22	>100	1	>100
7	>100	0.93	>100	0.93	67.3 ± 4.13	1.38	>100	0.93	>100	0.93	92.7 ± 5.00
8	>100	1	>100	1	>100	1	>100	1	>100	1	>100
9	>100	1	>100	1	96.5 ± 1.37	1.03	>100	1	>100	1	>100
Doxorubicin	0.7 ± 0.1	0.4	0.3 ± 0.1	1.0	1.6 ± 0.23	0.19	0.2 ± 0.09	1.5	1.1 ± 0.12	0.27	0.3 ± 0.1
Cisplatin	10.4 ± 0.9	0.6	6.7 ± 1.1	0.9	7.8 ± 0.98	0.81	3.2 ± 1.24	5.08	2.6 ± 0.15	2.42	6.3 ± 0.7

, and show that compounds 6–9 were not toxic to the normal keratinocyte HaCaT cell line in the range of the tested concentration. The analysis of the anticancer activity showed that the tested derivatives were mostly inactive (IC₅₀ > 100 µM). Dimer 6 had poor activity in relation to SW480, MDA-MB-231, and A-549 cell lines (IC₅₀ = 81–92 µM). Dimer 7, which contains two 1,8-dipyridothiazine units in its structure, was the most active in the group of tested compounds in relation to the MDA-MB-231 cell line (IC₅₀ = 67.3 µM). The results of the biological assay were very surprising when analyzed in relation to the previously obtained results of the anticancer activity of structural analogs 6a–9a, as shown in Figure 1 [25].

Lutidine derivatives 6a–9a showed activity in the range of IC₅₀ = 0.1–11 µM towards MCF 7 and IC₅₀ = 4.5–25.9 µM towards the SW480 cell lines [25]. This led us to consider the reason for the deactivation of the obtained dimers 6–9. To explain this phenomenon, quantum chemical calculations and preliminary molecular docking analyses were used.

3.5. Quantum Mechanical Calculations and Molecular Docking Study

In order to understand and explain the differences in the activity and inactivity of the dimers with an m-xylene and lutidine linker, measurements of the electron density of the selected dimer 8 containing 2,7-diazaphenothiazine in its structure and its analog with the lutidine 8a system were performed using the Gaussian program [32]. Dimer 8 was selected for quantum mechanical studies and molecular docking because it was the only one that showed inactivity against all of the tested cancer cell lines (IC₅₀ > 100 µM). A visualization of the obtained results is presented in Figure 2 and Figure 3 using the Avogadro program [33]. These measurements show that the presence of the lutidine system in dimer 8a causes a differentiation of the electron charges on the thiazine nitrogen atom in 1,4-thiazine rings directly connected to the CH₂ moiety (Figure 3). Additionally, the lutidine system contains a basic nitrogen atom with a lone electron pair, which causes the entire molecule to be enriched with electrons. This situation was not observed in dimer 8 with the m-xylene linker (Figure 2).

Additionally, bond length measurements were made, with particular attention paid to the connecting linkers. In the dimer 8a containing lutidine, there are CH₂-C_lutidine bonds with a length of 1.516 Å and a characteristic C_lutidine-N_lutidine bond with a length of 1.362 Å. These bonds are essentially shorter than the analogous bonds found in dimer 8 (CH₂-C_m-xylene with a length of 1.521 Å; C_m-xylene-C_m-xylene bond with a length of 1.405 Å) (Supplementary Materials File S1). The structural analysis carried out using quantum mechanical calculations to some extent explains the differences in the structure of isomeric dimers 8 and 8a, which had a direct impact on the affinity for protein molecular targets.

In the next step, in order to more deeply explain the weak activity of the dimers with the m-xylene system, preliminary molecular docking studies of dimer 8 and its lutidine derivative 8a were performed in relation to the histone deacetylase (HDAC4), indicated as a molecular target by the PASS program. The selected histone deacetylase (HDAC4) and the dimers 8 and 8a were prepared for docking according to the literature data (see Section 2.6) [34,35,36]. Two types of docking were performed using two kinds of software. The first one was performed using the DiffDock software Version 1.1.2 [37] with the default settings. The second docking was also performed using the AutoDock Vina software Version 4.2.6 [39], implemented on the TAMARIND BIO server, where the docking site was defined as the inhibitor-binding pocket in the crystallographic structure. This pocket was also detected during blind docking using DiffDock software when analyzing the selected ligands. The docking results are summarized in

Table 4. Docking results obtained from the DiffDock and AutoDock Vina tools.

Pose	DiffDock Confidence	Autodock Vina Binding Affinity [kcal/mol]	DiffDock Confidence	AutoDock Vina Binding Affinity [kcal/mol]
	8		8a
1	−1.26	−8.88	−1.38	−9.03
2	−1.34	−8.77	−1.42	−9.01
3	−1.46	−8.74	−1.51	−8.98
4	−1.61	−8.65	−1.59	−8.93
5	−1.91	−8.61	−1.60	−8.91
6	−2.06	−8.61	−1.73	−8.85
7	−2.18	−8.58	−1.93	−8.53
8	−2.38	−8.50	−1.98	−8.61
9	−3.25	−8.48	−2.08	−8.54
10	−3.67	-	−4.15	-
Average	−2.11	−8.65	−1.94	−8.82
Deviation	0.81	0.13	0.81	0.20

, which presents the confidence values of the DiffDock software, as well as the binding affinity calculated by the AutoDock software. The results indicate that both dimers 8 and 8a dock with a similar strength.

In a further stage of the docking study, an interaction analysis was performed for the best position from each complex using the Protein Ligand Interaction Profiler (PLIP) server [39] (

Table 5. Detected protein–ligand interactions for the analyzed complexes.

DiffDock	AutoDock Vina	DiffDock	AutoDock Vina
8		8a
Phe168, His198	Lys20, His21, Arg37, Ser123, Phe168	His158, Phe168, His198, Phe227	Arg37, Pro155, Asn119, Phe168, His198, His332

).

The analysis showed that the tested dimers 8 and 8a bind in different parts of the pocket, depending on the software used. Both docking programs indicated that dimer 8 had fewer binding sites than dimer 8a.

In the complexes obtained using the DiffDock server, the ligands bind near the zinc atom, Phe168 and His198 amino acids, while the complexes obtained using the AutoDock software in the left part of the pocket also interact with Phe168 (Figure 4).

The obtained results were also compared with the position of the trifluoromethyl ketone inhibitor in the structure of the tested protein. This inhibitor binds to the pocket into which test compounds 8 and 8a were docked using DiffDock software (Figure 5).

Analyzing the interactions of both dimers in this binding pocket, it can be noticed that dimer 8 is located at a considerable distance from the zinc atom and interacts only with Phe168 and His198 (Figure 6).

Meanwhile, the 8a dimer, having a lutidine system in its structure, is located closer to the zinc atom compared to its analog 8, and has four interactions with the following amino acids: Phe168, Phe277, His153, and His198 (Figure 7).

The performed quantum mechanical studies and preliminary molecular docking in relation to histone deacetylase explain the weak anticancer activity of the tested compounds compared to previously obtained active dimers with a lutidine linker.

4. Conclusions

The synthesis of dipyridothiazine dimers with an m-xylene moiety is presented in the light of a comparative analysis with their anticancer active analogs containing a lutidine linker in their structure. The structure of the new compounds was clearly confirmed by the NMR and HRMS spectroscopic techniques. The compounds were subjected to initial in silico screening using the Way2Drug server to determine the molecular targets and the likelihood of cytotoxicity against various types of cancer cells. In the next stage, cytotoxicity tests were performed against normal cell lines (HaCaT) and five cancer cell lines (SW480, SW620, MDA-MB-231, A-549, and LN-229), which showed weak anticancer activity and low cytotoxicity compared to the previously described structural analogs with the lutidine system. Quantum mechanical calculations of the electron density and bond length were performed to elucidate the structural differences that directly affected the conformation of the molecule, as well as interactions with the molecular targets. In addition, preliminary molecular docking studies were performed on histone deacetylase, showing different binding sites to the selected molecular target. Quantum mechanical calculations and molecular docking explained the anticancer activity research results to some extent.