Design, synthesis, and biological evaluation of novel thiazole derivatives as PI3K/mTOR dual inhibitors

Abstract

The development of anticancer drugs targeting both PI3K and mTOR pathways is recognized as a promising cancer therapeutic approach. In the current study, we designed and synthesized seventeen new thiazole compounds to investigate their effect on both PI3K and mTOR as well as their anti-apoptotic activity. All the synthesized thiazoles were investigated for their antiproliferative activity on a panel of 60 different cancer cell lines at the National Cancer Institute. Compounds 3b and 3e were selected for further investigation at five dose concentrations due to their effective growth inhibiting activity. Compounds 3b and 3e were further evaluated for their in vitro inhibitory activities against PI3Kα and mTOR compared to alpelisib and dactolisib, respectively as reference drugs. The inhibitory effect of compound 3b on PI3Kα was similar to alpelisib, but it showed weaker inhibitory activity on mTOR compared to dactolisib. Moreover, compound 3b exhibited significantly higher inhibitory activity compared to compound 3e against both PI3Kα and mTOR. The cell cycle analysis showed that compounds 3b and 3e induced G0–G1 phase cell cycle arrest in the leukemia HL-60(TB) cell line. Meanwhile, they significantly increased the total apoptotic activity which was supported by an increase in the level of caspase-3 in leukemia HL-60(TB) cell lines. Molecular docking experiments provided additional explanation for these results by demonstrating the ability of these derivatives to form a network of key interactions, known to be essential for PI3Kα/mTOR inhibitors. All these experimental results suggested that 3b and 3e are potential PI3Kα/mTOR dual inhibitors and could be considered promising lead compounds for the development of anticancer agents.

A series of new thiazole derivatives 3a–q were synthesized and tested against 60 cancer cell lines at the NCI, USA. Compound 3b showed the highest activity against PI3Kα with IC₅₀ = 0.086 ± 0.005 μM and IC₅₀ of 0.221 ± 0.014 μM against mTOR.

1. Introduction

Cancer is one of the major causes of mortality globally with an incidence of 18.1 million cancer cases around the world in 2020.¹ Cancer is a disease that is characterized by various types, among which is leukemia which affects the tissues in the body that generate blood (such as the lymphatic system and bone marrow). Leukemia contributed to approximately 2.5% and 3.1% of all new cancer incidence and mortality, with 311 594 deaths being reported in 2020.²

One of the most common etiologies of cancer, especially leukemia, is the aberrant upregulation of the PI3K/AKT/mTOR cascade.³ Phosphatidylinositol-3-kinases (PI3Ks) are lipid kinases that are responsible for the phosphorylation of the 3-position of phosphatidylinositol diphosphate (PIP2) to form phosphatidylinositol triphosphate (PIP3). PIP3 will phosphorylate and activate Akt protein which in turn activates mTOR leading to cell growth and proliferation. Excessive activity of this pathway causes extended survival of cells leading to tumor cells and cancer (Fig. 1). Therefore, the development of inhibitors targeting the PI3K/mTOR pathway has become a promising target for medicinal chemists around the world.^4–9 Dual PI3K/mTOR inhibitors have gained widespread interest because they can act both on PI3K and mTOR proteins, having several advantages such as better activity and less liability to drug resistance.¹⁰

Fig. 1. Targeting the PI3K/Akt/mTOR signaling pathway in cancer treatment.

Several inhibitors, in the last decade, have been reported as PI3K inhibitors or dual inhibitors of both PI3K and mTOR. Many of these inhibitors have entered clinical trials and many have been approved by the FDA (Food and Drug Administration) (Fig. 2). Structural similarities between the active site of PI3K and the catalytic domain of mTOR could be exploited to direct the design of dual PI3K/mTOR inhibitors.¹¹ Most of the PI3Kα inhibitors share similar structural and chemical features, i.e., a central heterocyclic ring, which binds to the hinge region in the active site, substituted by a hydrophobic group and a (hetero)aromatic cycle which binds to the affinity pocket and the ribose pocket, respectively^6,12 (Fig. 2). Considering the aforementioned information and focusing our research efforts in developing new anticancer agents with high efficacy,^13–21 target compounds were designed to retain a thiazole ring as the alpelisib drug. The thiazole ring is connected to both a hydrophobic group via a hydrazone group from one side and a substituted phenyl ring from the other side to maintain the pharmacophoric features of the dual inhibitors. Some of the hydrophobic groups that were linked to the hydrazine were chosen to mimic PI3K and PI3K/mTOR dual inhibitors; for example, substituted catechol derivatives were added to mimic the copanlisib drug, and the phenyl substituted diazolone ring to mimic the dactolisib drug. The hydrazone spacer was chosen due to the presence of hydrogen bond donors and acceptors as well as its flexible skeleton that makes it known for its superior anticancer activities.⁹

Fig. 2. Structures of A) PI3K inhibitors and B) PI3K/mTOR dual inhibitors and C) design of the target compounds.

In this article, we disclosed the synthesis of a series of new thiazole derivatives 3a–q. The cytotoxicity of each drug was evaluated using NCI-60 cell lines. The most potent derivatives were evaluated for their PI3Kα and mTOR inhibition activity as well as the caspase and cell cycle analysis. In addition, a molecular modelling study was performed to evaluate the binding interactions of the compounds in their active site and to computationally evaluate the ADME properties of the synthesized compounds.

2. Materials and methods

2.1. Chemistry

2.1.1. General procedures

The melting points, elemental analysis results, and spectral data were obtained as stated in our previous study.²² Details are in the ESI.† Compounds 1d–f^23,24 and 2a–g^25–28 were prepared according to the reported procedures. Compound 3q was previously reported as an antioxidant and a selective hMAO-B inhibitor.²⁹

2.1.2. The preparation of compounds 3a–q (general procedure)

Equimolar quantities of thiosemicarbazone derivatives 2a–g (0.01 mol) and the appropriate phenacyl bromide (0.01 mol) in absolute ethanol (20 mL) were refluxed for 8–12 h. The precipitate formed while hot was filtered, dried, and recrystallized from ethanol.

The detailed experimental data of all the synthesized derivatives are presented in the ESI.†

2.2. Biological evaluation

2.2.1. Biological assays

The biological assays were performed in compliance with previously published protocols and are included in the ESI.† Antiproliferative activity screening (60 cell lines) and 5 dose concentration assay by NCI,^30–33 MTT assay,³⁴in vitro enzymatic PI3Kα assay,¹⁷in vitro enzymatic mTOR assay,¹² cell cycle analysis,³⁵ apoptosis assay,³⁶ and caspase-3 enzyme assay³⁷ were also performed.

2.2.2. Computational study

2.2.2.1. Molecular modeling study

The 3D X-ray structure of the PI3Kalpha enzyme (PDB: 7PG6) and mTOR (PDB: 4JT6) were obtained from the protein databank (PDB) website, (https://www.rcsb.org/) at a resolution of 2.50 Å and 3.60 Å respectively. The docking studies were performed using Molecular Operating Environment computational software (MOE 2020.09; Chemical Computing Group, Canada). Initially, all the hydrogen atoms were added using the Protonate 3D algorithm where the protonation states of the amino acid residues were assigned, and the partial charges of atoms were added. In addition, the compounds were drawn using the builder tool and energy was minimized using the MMFF94x force field. The MOE induced-fit dock tool was used to dock the synthesized compounds into the active site. The selection of the final docked ligand–enzyme poses was according to the criteria of the binding energy score and combined with ligand–receptor interactions.

2.2.2.2. Prediction of physicochemical, pharmacokinetic, and ADME properties

The physicochemical descriptors of each of the synthesized compounds, the pharmacokinetic properties, the ADME, and the drug-like nature of all the compounds were calculated using the SwissADME web tool provided by the Swiss Institute of Bioinformatics (SIB). For calculation, structures were uploaded as SMILES notations.

2.2.2.3. Prediction of toxicity

The molecular structures of the most active compounds 3b and 3e were transformed into the SMILES. The SMILES of the compounds were then entered into the ProTox 3.0 website (https://tox.charite.de/protox3/) as input to predict the toxicity of the tested compounds.³⁸

2.2.3. Statistical analysis

Data are represented as mean ± SD. Significant differences between groups were analyzed by using Graphpad Prism 9.1.0. Differences were considered significant at p < 0.05.

3. Results and discussion

3.1. Chemistry

The general strategy for the synthesis of the target thiazole compounds 3a–q is outlined in Scheme 1. Thiosemicarbazide was reacted with an equimolar quantity of different aldehydes 1a–g in absolute ethanol with the addition of glacial acetic acid (few drops) under reflux to afford thiosemicarbazone derivatives 2a–g following a reported procedure.^25–28 Heterocyclization of thiosemicarbazone derivatives 2a–g with an equimolar quantity of appropriately substituted phenacyl bromide derivatives afforded the hydrazinyl-1,3-thiazole derivatives 3a–q as pure products in good to excellent yields. A variety of aldehydes including heterocyclic, aromatic, and polycyclic aromatic ones were used for the structural variation in designed thiazole derivatives 3a–q to determine the structure–activity relationship. According to the literature,^23,24 the substituted aldehydes utilized to prepare compounds 2d–f were prepared by reacting substituted 4-hydroxy benzaldehyde with benzyl chloride in anhydrous DMF with anhydrous K₂CO₃.

Scheme 1. Steps of the synthesis of the target thiazole compounds 3a–q.

The newly synthesized compounds were verified by elemental analyses, IR, ¹H NMR, and ¹³C NMR spectral data. ¹H NMR spectra of compounds 3a–q showed sharp singlets at δ 7.88–8.10 ppm indicating the presence of characteristic azomethine (CH N) protons. Additionally, a D₂O exchangeable singlet signal between δ 11.82 and 12.65 ppm showed that the hydrazide NH protons were present. Their ¹³C NMR spectra fully agreed with the suggested structures and the evidence for thiazole ring formation was confirmed by a characteristic C₂ signal of the thiazole ring that appeared at δ 164.6–172.7 ppm.

The IR spectrum of compounds 3a and 3b exhibited a characteristic absorption band at 3313 and 3479 cm⁻¹ respectively due to the stretching of the –OH group. ¹H NMR spectroscopy disclosed ethyl proton triplet and quartet signals, confirming the ethoxy function. ¹³C NMR supported the presence of the ethoxy group by revealing two signals at δ 64.3 and 15.0 ppm.

¹H NMR spectral data of derivatives 3f and 3g were also consistent with the assigned structures, and C–CH₃ and N–CH₃ protons of the pyrazolone moiety appeared in the region of δ 2.61–3.26 ppm. Meanwhile, IR spectra were very informative, with a significant band at 1647 cm⁻¹ due to the pyrazolone carbonyl group stretching. Moreover, ¹³C NMR spectra exhibited signals for the aliphatic carbon atoms in the expected regions and the signals of carbonyl groups resonated at δ 164.0 and 161.8 ppm, respectively. ¹H NMR spectra of compounds 3h–p showed a singlet signal at δ 5.13–5.16 ppm suggesting –OCH₂ protons and this was substantiated by ¹³C NMR spectra where the carbon directly linked with oxygen was found at δ 69.8–70.4 ppm.

3.2. Biological evaluation

3.2.1. Antiproliferative activity against sixty human cancer cell lines

Seventeen thiazole derivatives were evaluated by the Developmental Therapeutic Program NCI, USA,^30–33 for their cytotoxic potential utilizing 60 distinct human tumor cell lines, including leukemia, melanoma, ovarian, renal, prostate, and breast cancers as well as tumors of the non-small cell (NSC) lung, colon, central nervous system (CNS). Single dose screening (10 μM) was performed on all compounds and the data of growth inhibition percentages (GI%) are presented in Table S1 in the ESI.† The results show that compounds 3b with the 2-ethoxyphenol group and 3e with the 3-chloro-4-nitrophenyl ring exhibited excellent, broad-spectrum anticancer activity against most of the examined cell lines with mean GI above 60% (Fig. 3) and they were subsequently evaluated at five concentrations. Compound 3b was the most potent against all tested compounds and it exhibited potent cytotoxic efficacy against 20 cancer cells having GI% values between 76.97 and 99.95%. It also demonstrated a lethal effect (GI above 100%) against 36 cell lines. Compound 3e showed a good-to-significant antiproliferative efficacy against the majority of leukemia lines with GI% between 81.15% and 86.93%. It showed a lethal effect on RPMI-8226 and leukemia HL-60(TB) cell lines. Additionally, compound 3e showed strong antiproliferative effects in 11 cell lines with growth inhibitions between 74.32 and 99.19%. It also showed a lethal effect against 12 cell lines. Compounds 3a, 3d, 3o, and 3p exhibited potent-to-moderate cytotoxic activity with mean GI above 30% (36.86%, 36.04%, 37.06%, and 31.83%, respectively) (Fig. 3). Compound 3a displayed potent cytotoxicity with GI% between 73.44 and 89.79% against 5 cell lines and it showed moderate inhibition against 28 cell lines with GI% in the range of 30.88–67.47%. 3a has a lethal effect on colon cancer HCC-2998. Moreover, compound 3d displayed excellent cytotoxicity against 7 tumor lines with GI% of 75.71–97.95% and a lethal activity against 4 cell lines. Compound 3o demonstrated promising GI in non-small cell lung (A549/ATCC and NCI-H460) and colon (HCT-116 and SW-620) cell lines within a GI% range of 76.18–88.78%. Moreover, it demonstrated moderate inhibition versus 32 tumor lines within the GI range of 31.25–68.17%. Also, compound 3p established promising GI in non-small cell lung (A549/ATCC and NCI-H460) and colon (HCT-116, HT29, and SW-620) cell lines within a GI% range of 70.93–78.60%. It displayed moderate inhibition against 25 cell lines within a GI range of 30.01 to 65.84%.

Fig. 3. Graphical representation of mean GI percentages of the synthesized derivatives 3a–q.

For the rest of the series, compounds 3c, 3f–n, and 3q exhibited moderate to low cytotoxic activity with a mean GI% range of 9.82–28.81% (Fig. 3). Compound 3c displayed potent cytotoxic activity against leukemia (RPMI-8226), colon (KM12), ovarian (OVCAR-3), lung cancer (NCI-H322M), and breast (MCF7) cancer cell lines in the GI% range of 83.11–96.18. It also showed a lethal effect against 4 cancer cell lines. However, the antipyrine derivative 3 demonstrated a moderate level of efficacy against leukemia K-562, NSC NCI-H522, CNS SF-539, ovarian OVCAR-4, renal (786-0 and ACHN) and breast (MDA-MB-231/ATCC, HS 578 T, T-47D, and MDA-MB-468) cancer cell lines with GI% ranging from 30.32–51.18%. Unfortunately, compound 3g did not display significant activity against most tested cell lines. Meanwhile 4-(benzyloxy) benzylidene derivative 3h demonstrated substantial antiproliferative activity against the ACHN renal cell line with 79.82% inhibition and modest antiproliferative efficacy against eight cancer cell lines with a cell growth inhibition range of 34.93–50.17%. Moreover, compound 3i demonstrated moderate GI% in sixteen cancer cell lines within a range of 31.12–60.14%. Also, compound 3j demonstrated moderate GI% in nineteen cell lines within a range of 31.89–54.03%. Meanwhile, compound 3k showed good antiproliferative action against all leukemia cell types with inhibition ranging from 51.06 to 65.07%. In addition, it displayed moderate cytotoxicity against eighteen tested panels, within a GI range of 30.65–68.51%. The majority of the tested cell lines showed moderate to weak activity for compound 3l, with the best inhibition (45.02%) induced against the MCF7 breast cancer cell line. Additionally, compound 3m showed mild to moderate action against most of the cell lines tested with maximum inhibition (48.12%) against the UO-31 renal cell line. Meanwhile, compound 3n showed significant antiproliferative activity against only the RXF 393 renal cell line with 79.23% inhibition. Moreover, it showed modest antiproliferative efficacy against 17 tested panels except for melanoma where it inhibited cell proliferation in the range of 31.73–65.91%. Finally, compound 3q showed relatively moderate efficacy against only nine cell lines within the GI% range of 30.45% to 57.12%.

Based on the mean growth inhibition, the following structure–activity relationship was observed. Results in Table S1† showed the impact of varying the aryl moiety attached to the hydrazone group from one side of the thiazole ring. The order of reactivity noted is as follows: 2-ethoxyphenol > 3-chloro-4-nitrophenyl ring > 4-(benzyloxy)-3-ethoxyphenyl > 4-(benzyloxy)-3-methoxyphenyl derivatives. Meanwhile, the incorporation of 4-(benzyloxy)phenyl, antipyrine and 1,3-benzodioxole rings as aryl moieties did not remarkably improve the cytotoxic activity. Generally, the substitution with the nitro group on position 3 of the distal phenyl ring displayed the best activity among the 2-ethoxy phenol, antipyrine, 4-(benzyloxy)benzylidene and 4-(benzyloxy)-3-methoxybenzylidene series. On the other hand, the substitution with the chloro group on position 4 of the distal phenyl ring displayed good activity among derivatives with 3-chloro-4-nitrophenyl, and 4-(benzyloxy)-3-ethoxybenzylidene aryl rings. It is generally noted that derivatives with a terminal unsubstituted phenyl ring were less active compared to derivatives with a substituted phenyl ring except for 3j. Moreover, results showed that the thiazole derivative connected to the 2-ethoxyphenol moiety attached to the hydrazone group from one side and the 3-nitrophenyl ring from the other side; compound 3b exerted the highest cytotoxic potential with 3-fold higher activity compared to its unsubstituted counterpart 3a. Also, the thiazole derivative connected to the 3-chloro-4-nitrophenyl ring attached to the hydrazone group from one side and the 4-chlorophenyl ring from the other side; 3e exerted potent cytotoxic potential with 3-fold higher activity compared to its unsubstituted counterpart 3c. Regarding the 4-(benzyloxy)benzylidene derivatives, compound 3i showed 1.4-fold higher activity compared to its unsubstituted counterpart 3h. Among the 4-(benzyloxy)-3-methoxybenzylidene derivatives 3j–3l, the 3-nitro substituted derivative 3k was more active than the 4-chloro substituted or unsubstituted counterparts 3l and 3j, respectively. Meanwhile, in the 4-(benzyloxy)-3-ethoxybenzylidene series 3m–3p, 3o substituted with 4-Cl on the distal phenyl ring displayed 2.6-fold higher activity compared to the unsubstituted counterpart 3m. Also, 3p substituted with 4-Br on the distal phenyl ring displayed 2.2-fold higher activity compared to the unsubstituted counterpart 3m. Finally, 3n substituted with 3-nitro on the distal phenyl ring displayed 1.6-fold higher activity compared to its unsubstituted counterpart 3m. From the above results, we can note that the order of reactivity among 4-(benzyloxy)-3-substituted benzylidene derivatives is as follows: ethoxy > methoxy > unsubstituted derivatives. Finally, it was observed that halogen containing derivative 3l was the least active among the 4-(benzyloxy)-3-methoxybenzylidene series 3k–3l while the opposite effect was observed among the 4-(benzyloxy)-3-ethoxybenzylidene series 3m–3p where the halogen derivatives 3o and 3p were the most active (Fig. 4).

Fig. 4. Structure–activity relationship (SAR) studies of the synthesized compounds based on mean growth inhibition.

3.2.2. Calculating the GI₅₀, TGI, and LC₅₀ values of compounds 3b and 3e using the NCI-60 cell line panel

As shown from the above results, significant antiproliferative activity was demonstrated by two compounds (3b and 3e) at a 10 μM dose with respective mean growth rates of 111.56% and 65.96%. They were then assessed at five different concentrations (0.01–100 μM) to determine the dose–response behavior and compute their GI₅₀ (dosage inhibiting 50% of growth relative to the control), TGI (dosage completely inhibiting growth), and LC₅₀ (dosage killing 50% of the initial cell count). Table 1 displays values of GI₅₀, TGI, and LC₅₀ for compounds 3b and 3e. Compound 3b displayed strong antiproliferative activity against most of the studied tumor lines, based on the findings of five dose screening, with GI₅₀ readings between 1.71 and 4.32 μM. As for compound 3e, it displayed strong antiproliferative activity against 25 tumor cell lines with GI₅₀ values ranging from 1.54 to 5.86 μM. Aside from that, compound 3b demonstrated strong cytostatic action (range of TGI: 3.23–9.90 μM) against 32 cancer cell lines from each studied cancer subpanel, except the prostate subpanel. Compound 3b also showed good to weak cytostatic efficacy toward the remaining cancer cell lines with a 10.90–44.80 μM TGI range. Moreover, compound 3e had strong cytostatic action (TGI range of 3.01–7.43 μM) toward 11 cancer cell lines from all examined cancer subpanels, except renal cancer, NSC lung cancer, melanoma, and CNS subpanels. Additionally, with a TGI of 10.80 to 98.30 μM, it demonstrated cytostatic efficacy varying in effectiveness against the remainder of the cancer cell lines from good to weak. The MOLT-4 leukemia cancer cell line was the most sensitive to compound 3b, which showed values of 1.74 and 5.94 μM for GI₅₀ and TGI, respectively. Furthermore, the leukemia HL-60(TB) cell line was effectively and selectively inhibited by 3b with a GI₅₀ of 2.32 μM and a TGI of 7.84 μM. 3b demonstrated strong and selective antiproliferative action against the NSC lung carcinoma NCI-H226 and NCI-H460 cell lines with GI₅₀ values of 2.36 and 2.40 μM and TGI = 6.22 and 5.81 μM respectively along with LC₅₀ values above 100 μM demonstrating their safety profile for all these cell lines. Compound 3e was discovered to be especially sensitive to the leukemia HL-60(TB) and RPMI-8226 cell lines as well as the breast HS 578T cell line. Their GI₅₀ values were 2.29, 2.31, and 2.36 μM and their TGI values were 7.43, 6.29, and 6.14 μM, respectively, with LC₅₀ values above 100 μM. Compounds 3b and 3e demonstrated significant differences in their cytotoxic indicator (LC₅₀) and cytostatic markers (GI₅₀ and TGI) against the leukemia cancer HL-60(TB) cell line, suggesting a broad therapeutic index.

Measurement of GI₅₀, TGI, and LC₅₀ values for compounds 3b and 3e against the 60-cell line panel of the NCI.

Panel/cell line	Compounds (μM)
	3b			3e
	GI50	TGI	LC50	GI50	TGI	LC50
Leukemia
CCRF-CEM	2.03	29.60	>100	3.62	52.90	>100
HL-60(TB)	2.32	7.84	>100	2.29	7.43	>100
K-562	3.05	34.30	>100	5.86	66.40	>100
MOLT-4	1.74	5.94	>100	3.43	24.10	>100
RPMI-8226	3.03	>100	>100	2.31	6.29	>100
SR	3.07	>100	>100	NT	NT	NT

Non-small cell lung cancer
A549/ATCC	3.45	14.10	80.60	78.40	>100	>100
EKVX	3.17	16.40	>100	10.30	22.90	51.20
HOP-62	2.51	6.92	43.30	8.87	22.40	52.50
HOP-92	3.20	30.90	>100	2.59	10.80	60.20
NCI-H226	2.36	6.22	>100	4.21	32.60	>100
NCI-H23	1.99	5.14	25.70	12.50	40.90	>100
NCI-H322M	4.32	16.90	50.50	3.23	13.70	37.10
NCI-H460	2.40	5.81	>100	13.60	26.50	51.50
NCI-H522	1.85	4.21	9.58	9.45	67.60	>100

Colon cancer
COLO 205	16.40	38.70	91.20	15.40	29.70	57.20
HCC-2998	3.43	19.80	>100	1.87	3.55	6.72
HCT-116	2.34	5.83	36.70	16.30	36.60	82.10
HCT-15	2.86	11.80	72.60	13.10	42.10	>100
HT29	33.70	>100	>100	3.60	15.40	>100
KM12	1.94	3.63	6.81	1.88	3.58	6.80
SW-620	3.53	11.60	97.90	18.30	43.10	>100

CNS cancer
SF-268	3.45	27.40	>100	8.44	31.80	>100
SF-295	2.00	4.28	9.16	14.90	28.30	53.80
SF-539	1.78	4.15	9.68	16.30	31.10	59.30
SNB-19	3.35	20.80	>100	10.50	24.20	55.70
SNB-75	2.72	13.10	>100	14.80	40.10	>100
U251	3.08	14.10	>100	4.37	17.00	46.10

Melanoma
LOX IMVI	1.73	3.59	7.46	3.91	>100	>100
M14	1.85	3.89	8.19	17.00	45.60	>100
MDA-MB-435	2.70	9.14	43.00	23.20	98.30	>100
SK-MEL-2	2.48	6.75	26.40	24.00	81.50	>100
SK-MEL-28	2.27	5.97	27.40	17.20	46.40	>100
SK-MEL-5	2.73	9.14	31.50	10.30	24.50	58.00
UACC-257	1.71	3.23	6.08	7.52	22.00	51.60
UACC-62	2.39	6.56	24.90	11.10	26.00	60.90

Ovarian cancer
IGROV1	2.46	7.23	91.80	13.60	37.40	>100
OVCAR-3	2.31	4.43	8.49	1.73	3.13	5.69
OVCAR-4	2.41	8.05	>100	8.25	35.90	>100
OVCAR-5	3.79	16.20	51.30	23.00	78.20	>100
OVCAR-8	2.84	9.63	39.60	14.90	54.10	>100
NCI/ADR-RES	2.72	8.59	68.30	16.80	41.50	>100
SK-OV-3	3.36	27.80	>100	21.80	65.00	>100

Renal cancer
786-0	2.75	8.68	81.80	19.80	61.30	>100
A498	20.20	44.80	99.70	14.80	39.60	>100
ACHN	3.22	11.60	61.40	15.10	28.40	53.40
CAKI-1	2.18	6.88	26.80	14.60	32.10	70.60
RXF 393	1.89	4.82	>100	11.90	30.60	78.70
SN12C	2.41	9.90	>100	15.10	61.50	>100
TK-10	4.42	16.10	53.00	28.10	84.00	>100
UO-31	2.29	8.85	>100	8.46	20.70	45.50

Prostate cancer
PC-3	2.54	11.40	>100	1.91	4.46	12.50
DU-145	3.43	11.60	38.90	1.63	3.07	5.81

Breast cancer
MCF7	2.86	10.90	43.20	1.79	3.81	8.11
MDA-MB-231/ATCC	1.78	3.88	8.48	17.40	42.60	>100
HS 578 T	3.38	39.20	>100	2.36	6.14	>100
BT-549	2.84	8.12	>100	14.50	40.40	>100
T-47D	3.69	16.30	>100	1.80	3.81	8.05
MDA-MB-468	2.17	4.61	9.79	1.54	3.01	5.89

3.2.3. MTT assay and selectivity index (SI) calculation

The selectivity index (SI) was determined to evaluate the toxicity of the most potent derivatives, 3b and 3e, on normal cells, with the results presented in Table 2. In this selectivity study, IC₅₀ values were measured against primary CD8+ cytotoxic T cells PCS-800-017, using doxorubicin as a positive control. The SI was calculated as the ratio of cytotoxicity (IC₅₀) on normal lymphocytes PCS-800-017 to that on the leukemia cell line HL-60(TB). Compared to doxorubicin, which had an IC₅₀ of 12.52 ± 0.38 μM against PCS-800-017 cells and an SI of 27.82, compounds 3b and 3e showed IC₅₀ values of 28.56 ± 0.94 and 22.67 ± 0.74 μM, respectively, against normal lymphocytes PCS-800-017, demonstrating good selectivity for cancer cells with SI values of 12.31 and 9.90, respectively.

In vitro antiproliferative assessment of compounds 3b and 3e against leukemia HL-60(TB) and normal lymphocytes PCS-800-017 and selectivity index.

Compound	HL-60(TB)	PCS-800-017	Selectivity index
	IC50 μM ± SDa
3b	2.32	28.56 ± 0.94	12.31
3e	2.29	22.67 ± 0.74	9.90
Doxorubicin	0.45	12.52 ± 0.38	27.82

^aIC₅₀ values are presented as means of three independent experiments.

3.2.4. In vitro enzymatic dual PI3Kα/mTOR inhibitory assay

Developing dual PI3K/mTOR inhibitors could provide an ideal solution for the drug resistance and of a single inhibitor in cancer chemotherapy. To further understand the mechanism of cytotoxicity of the most potent derivatives 3b and 3e, the target derivatives were assessed in vitro for their inhibitory activities against PI3Kα and mTOR in comparison to alpelisib which is an FDA-approved PI3K inhibitor with selective activity against PI3Kα³⁹ and dactolisib, a marketed dual PI3K/mTOR inhibitor known for its broader inhibition across the inhibition of mTOR,⁴⁰ respectively, as reference drugs, and the IC₅₀ values are presented in Fig. 5. The enzyme inhibition rate/compound concentration curves are presented in the ESI.† The inhibitory activity of compound 3b against PI3Kα was comparable to the inhibitory activity of alpelisib with IC₅₀ values of 0.086 ± 0.005 and 0.087 ± 0.005 μM, respectively. However, the inhibitory activity of compound 3b against mTOR was lower than that of dactolisib with IC₅₀ values of 0.221 ± 0.014 μM and 0.119 ± 0.014 μM, respectively. The inhibitory activity of compound 3b was significantly higher than the inhibitory activity of compound 3e against both PI3Kα and mTOR with IC₅₀ values of 0.086 ± 0.005 and 0.221 ± 0.014 μM, respectively, for compound 3b and IC₅₀ = 0.148 ± 0.008 and 0.896 ± 0.055 μM, respectively for compound 3e. According to these findings, compound 3b can be considered as a potential PI3Kα/mTOR dual inhibitor and an antiproliferative drug.

Fig. 5. Graphs illustrating the IC₅₀ inhibition of compounds 3b and 3e against A) PI3Kα with the reference drug alpelisib and B) mTOR with dactolisib as a reference drug. Data are expressed as mean ± SE. Statistical significance was evaluated using one-way ANOVA and Tukey's multiple comparison test (*p < 0.05, ****p < 0.0001).

3.2.5. Cell cycle analysis

The analysis of the cell cycle distribution on the leukemia HL-60(TB) cell line was conducted to reveal the mechanism of compounds 3b and 3e in inhibiting cell proliferation. Using DNA flow cytometry analysis, we were able to determine the cell cycle distribution phases. HL-60(TB) cells were exposed to 3b and 3e derivatives at the respective GI₅₀ doses of 2.32 and 2.29 μM throughout a 24 h period to observe the effect of these compounds in various phases of the cell cycle. The findings of these studies are presented in Fig. 6. Both compounds 3b and 3e demonstrated a non-significant decrease in the percentage of cells in the S phase with 2.5% and 7.3%, respectively. However, they showed a significant decrease in the percentage of cells at the G2/M phase with a respective decrease of 63.5% and 12.8%, compared with the (DMSO) control. Furthermore, compounds 3b and 3e significantly increased the percentage of cells by 1.28- and 1.13-fold, respectively, in the G0–G1 phase. According to the results, the target derivatives 3b and 3e reduced cellular proliferation by producing cell cycle arrest in HL-60(TB) cells in the G0–G1 phase and inhibiting the proliferation of leukemia HL-60(TB) cells.

Fig. 6. Cell cycle arrest of compounds 3b and 3e in HL-60(TB) cells. (A) 3b treated cells; (B) 3e treated cells; (C) control HL-60(TB); (D) graphical display of the cell cycle distribution in both control and treated cells. Data are expressed as mean ± SE. Statistical significance was analyzed using one-way ANOVA and Dunnett's multiple comparison test (*p < 0.05, ***p < 0.001, ****p < 0.0001).

3.2.6. Apoptosis assay

In the present study, a dual-staining assay utilizing Annexin V and propidium iodide (PI) was applied to investigate the effect of compounds 3b and 3e on the apoptotic process of cells.^41,42 An apoptosis assay was conducted using HL-60(TB) leukemia cell lines, and the morphological indicators of apoptosis for the leukemia cells were analyzed before and during the treatment with compounds 3b and 3e for 24 h. Based on the results presented in Table 3 and Fig. 7, compounds 3b and 3e exposed an elevation in the Annexin V apoptotic cell percentage in the early apoptotic phase (from 0.61% to 31.19% and 23.62%, respectively) and the late apoptotic phase (from 0.28% to 16.27% and 17.92%, respectively). Compounds 3b and 3e induced an elevation of total apoptosis of 23.07- and 19.90-fold, respectively, compared to the control (DMSO).

Comparison of the distribution of apoptotic cells in HL-60(TB) cells subjected to compounds 3b and 3e to a control group using the Annexin V/PI dual staining test.

Comp.	Apoptosis			Necrosis
	Total	Early	Late
3b	52.39	31.19	16.27	4.93
3e	45.18	23.62	17.92	3.64
Cont. HL-60(TB)	2.27	0.61	0.28	1.38

Fig. 7. Impact of compounds 3b and 3e on the proportion of Annexin V positive staining in HL-60(TB) cells. (A) 3b treated cells; (B) 3e treated cells; (C) control HL-60(TB); (D) the graph represents the percentage of apoptotic and necrotic cells in both the control and treated cells.

3.2.7. Evaluation of caspase-3 activation

Caspases regulate the transmission of signals that start the process of programmed cell death or apoptosis. Effector caspases help tear down cells by cleaving structural proteins. Caspase 3 activation is very crucial to start the process of apoptosis.⁴³ Compounds 3b and 3e were tested for their capacity to induce apoptosis by assessing caspase-3 activation in the leukemia HL-60 (TB) cell line, in comparison with doxorubicin, a widely recognized chemotherapy drug that triggers apoptosis by activating caspase-3,⁴⁴ as the reference drug. When the results were compared to the untreated HL-60(TB) cells which scored a caspase-3 level of 109.5 ± 11.2 pg mL⁻¹, it was clear that compounds 3b caused a significant 4.61-fold increase in the level of caspase-3 (504.9 ± 15.7 pg mL⁻¹), compound 3e also caused a significant 3.48-fold increase in the level of caspase-3 (381.5 ± 11.2 pg mL⁻¹), and doxorubicin caused a significant 4.99-fold increase in the level of caspase-3 (546.6 ± 8.48 pg mL⁻¹). It was also clear that the effect of compound 3b on the caspase-3 level was comparable to that of doxorubicin (Fig. 8).

Fig. 8. The evaluation of caspase-3 in HL-60(TB) leukemia cells, comparison between compounds 3b, 3e, and doxorubicin versus untreated cells. Data are expressed as mean ± SE. Statistical significance was evaluated using one-way ANOVA and Tukey's multiple comparison test (*p < 0.05, ****p < 0.0001).

3.3. Computational study

3.3.1. Molecular docking study

Docking analysis was conducted using MOE (2020.09) to explore the potential binding mechanisms of the target compounds 3a–q with the active sites of PI3Kα and mTOR. The protein crystal structures of PI3Kα (PDB: 7PG6)⁴⁵ and mTOR (PDB: 4JT6)⁴⁶ were used and the results are presented in Table 4.

The docking score and Lipinski parameters of the most active compounds 3b and 3e.

Compound	Docking score with the PI3Kα active site (kcal mol−1)	Interactions with the PI3Kα active site	Docking score with the mTOR active site (kcal mol−1)	Interactions with the mTOR active site
3b	−6.9612	HB (Val851)	−7.7785	HB (Gly2238)
		HB (Arg770)		HB (Thr2245)
		HB (Asp933)		Arene–H (Met2345)
				Arene–H (Ile2356)
3e	−6.4525	HB (Val851)	−7.9819	HB (Val2240)
		HB (Arg770)		Arene–H (Glu2190)
		Arene–H (Tyr836)		Arene–H (Met2345)
		Arene–H (Val850)
		Arene–H (Gln859)
Alpelisib	−8.1121	HB (Ser854)
		HB (Gln859)
		HB (Lys802)
		HB (Val851)
PI-103			−7.2661	HB (Val2240)
				Arene–H (Trp2239)

Redocking the co-crystallized alpelisib in the PI3Kα active site served to validate the docking process. With the RMSD between the docked and the modeled alpelisib falling inside the cutoff limit, the docking procedure was able to replicate the pose of alpelisib. (Fig. 9). The structure of alpelisib consists of the 2-aminothiazole ring as a center ring attached to both the prolinamide ring and the substituted pyridine ring. The 2-amino thiazole ring is embedded in the hinge area of the active site and showed two hydrogen bonds with Val851. The prolinamide moiety fits in the ribose region of the active site and interacts with Ser854 and Gln859 via its amido group. Finally, the substituted pyridine is deeply embedded in the affinity region and the CF₃ group showed hydrogen bond interaction with the Lys802 residue of the enzyme.

Fig. 9. A) 2D diagram of alpelisib binding with the binding pocket; B) 3D of both the co-crystallized alpelisib inhibitor (grey) and its re-docked pose (magenta) in the active site of PI3Kα showing RMSD (0.907 Å) within the acceptable range.

The same docking procedure was then used for the synthesized target compounds to estimate their interactions with the PI3K binding site. The docking of the synthesized compound 3b (Fig. 10) displayed a similar binding mode with the active site as alpelisib. The nitro phenyl moiety of compound 3b inhabits the same space as the prolinamide ring and was able to form a hydrogen bond with Arg770.

Fig. 10. The binding modes of the compounds 3b and 3e in the ATP binding domain of PI3Kα. A) 2D diagram and B) 3D illustration of molecular docking of compound 3b in the PI3Kα binding pocket. C) 2D diagram and D) 3D illustration of molecular docking of compound 3e in the PI3Kα active site.

Furthermore, the thiazole ring interacts with Val851 by H-bonding, just like alpelisib does, increasing the stability of the drug conformation in the active site. Lastly, the p-hydroxyl group formed a hydrogen bond with Asp933. Compound 3e with the m-nitro phenyl substitution displayed higher interactions with the nitro group accepting a hydrogen bond from Arg770 and the phenyl ring interacts with Gln859 via the arene–H interaction. Further, the amino thiazole ring interacted with Val851 via a hydrogen bond and with Val850 and Tyr836 via arene–H interactions (Fig. 10). In addition, the docking of compounds 3b and 3e showed high binding to the mTOR active site compared to PI-103, a dual mTOR/PI3K inhibitor. The nitro group and the thiazole of compound 3b interact with Thr2245 and Gly2238 by a hydrogen bond, respectively. The 2 phenyl rings interact with Met2345 and Ile2356 via 2 arene–H interactions (Fig. 11). Compound 3e showed hydrogen bond interaction with Val2240 via the nitro group and 2 arene–H interactions with Glu2190 and Met2345 (Fig. 11) (all docking poses and results of the compounds 3a, 3c, 3d, 3f–q with PI3K and mTOR are included in ESI† Table S2, Fig. S1 and S2). In general, the results of the docking study showed that compounds 3b and 3e could interact with the active binding site of PI3K and mTOR and may be potential PI3K/mTOR dual inhibitors.

Fig. 11. The interactions of the compounds 3b and 3e in the mTOR active site A) 2D diagram and B) 3D illustration of molecular docking of compound 3b in the mTOR binding pocket. C) 2D diagram and D) 3D illustration of molecular docking of compound 3e in the mTOR active site.

3.3.2. Prediction of physicochemical, pharmacokinetic, and ADME properties

Additionally, several physicochemical and ADME parameters were predicted using SwissADME, a free online tool for pharmacokinetic analysis,⁴⁷ for all synthesized compounds. All compounds were tested for their compliance with Lipinski's rule of five,⁴⁸ obeying at least three of the four criteria is considered to adhere to Lipinski's rule. The compounds showed variable permeability based on gastrointestinal absorption (GI) and poor blood–brain barrier permeation⁴⁹ (Table S3, ESI†). Concerning oral bioavailability, the probability of oral bioavailability score of 0.55 is expected for all compounds. According to the bioavailability radar map (Fig. 12), compounds 3b and 3e showed reasonable physicochemical features and promising pharmacokinetic properties, along with high in vitro efficacy and high binding, and could be considered as potential candidates for further development.

Fig. 12. Bioavailability radar map of alpelisib, dactolisib, 3b and 3e. The pink area represents the optimal range for each property (lipophilicity: XLOGP3 between −0.7 and +5.0, size: MW between 150 and 500 g mol⁻¹, polarity: TPSA between 20 and 130 Ų, solubility: log S not higher than 6, saturation: fraction of carbons in the sp³ hybridization not less than 0.25, and flexibility: no more than 9 rotatable bonds).

3.3.3. Prediction of toxicity

As part of the process of developing new drugs, predicting the toxicities of compounds is critical. The tool is valuable in early-stage drug development and safety testing, reducing the reliance on animal testing. The toxicity of compounds 3b and 3e was predicted using ProTox 3.0 (online computational toxicology tool). ProTox 3.0 combines molecular similarity and most frequent features to predict the toxicity endpoints such as acute toxicity, organ toxicity, toxicological endpoints, metabolism, adverse outcome (Tox21) pathways and toxicity targets. Compounds 3b and 3e showed safety profile prediction against 30 different potential toxicities (Table S4, ESI†). The safety profile of compounds 3b and 3e encouraged our team to further optimize these compounds.

4. Conclusion

The current study described the design and synthesis of a series of new thiazole derivatives 3a–q as PI3K and mTOR dual inhibitors. Compounds 3b and 3e showed broad and strong antiproliferative activity on numerous cell lines with GI₅₀ values in the micromolar range of 1.71–4.32 μM and 1.54–5.86 μM, respectively. Compound 3b displayed almost comparable in vitro inhibitory activity (IC₅₀ = 0.086 ± 0.005 μM) compared to alpelisib against PI3Kα, while it showed less inhibitory activity compared to dactolisib against mTOR. Meanwhile, compound 3e showed lower inhibitory activity than 3b against both PI3Kα and mTOR with IC₅₀ = 0.148 ± 0.008 and 0.896 ± 0.055 μM, respectively. Results of cell cycle analysis showed that 3b and 3e caused an arrest in the cell cycle proliferation of HL-60(TB) leukemia cells at the G0–G1 phase. The most potent compounds 3b and 3e induced elevation of 23.07- and 19.90-fold in total apoptosis in HL-60(TB) cells as shown by Annexin V/PI assay. Compounds 3b and 3e elevated the level of apoptotic caspases-3 by 4.61- (3b) and 3.48-fold (3e). Molecular docking study showed that compounds 3b and 3e could interact with the active binding site of PI3K and mTOR. Finally, based on the bioavailability radar map, compounds 3b and 3e have the best in vitro efficacy, high binding, and reasonable physicochemical features, suggesting them as potential PI3K/mTOR dual inhibitors.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

The authors have clearly stated that they do not have any conflicts of interest. The authors are solely responsible for the content and writing of this manuscript.