A new class of anticancer activity with computational studies for a novel bioactive aminophosphonates based on pyrazole moiety

The present study involves synthesis a new series of α-aminophosphonates 2a-f and 4a-d derivatives in good yield with a simple workup via Kabachnik-Fields reaction in the presence of lithium perchlorate as Lewis acid catalyst. All the newly synthesized compounds were confirmed using various physical, spectroscopic, and analytical data. The in vitro anticancer activities of each compound were evaluated against colorectal carcinoma Colon cancer (HCT-116) and Epdermoid carcinoma (HEP2) and also Human lung fibroblast normal cell line (WI38) compared with Doxorubicin. The results showed that Compounds 2a, 4b and 4d exhibited more potent inhibitory activity for Epdermoid Carcinoma (HEP2) compared with doxorubicin. For colon carcinoma cells (HCT-116) Compounds 2a, 2d and 4b gave the strongest activity among all compounds compared with doxorubicin. Moreover, all designed structures were docked into the active site of VEGFR2 and FGFR1 proteins. The result reveals that compound 2b and have the strongest inhibitory activity of the VEGFR2 and FGFR1 proteins indicating that these substances might conceivably operate as VEGFR2 and FGFR1 inhibitors and hence might take role in anticancer activities with various binding interactions. The 3D-QSAR models produced strong statistical results since they were defined by PLS factors 4 and confirmed by parameters as R2, R2 CV, Stability, F-value, P-value, RMSE, Q2, and Pearson-r.

www.nature.com/scientificreports/normal cell line (WI38) besides different tumor cell lines.For the moment, researchers have been drawn to the design of more potent pyrazole derivatives having great diversity of biological activity (Fig. 1b) 14 .Moreover, the literature survey results introduce novel pyrazole-based moieties as dihydro-pyrano-pyrazole and pyrazolopyrimidine derivatives with 2-bromophenyl moiety as promising scaffolds to produce potent VEGFR-2 inhibitors.Additionally, over the last decade, the family of the fibroblast growth factor receptors (FGFRs) has become an attractive validated therapeutic target notably in cancer diseases.A series of novel pyrazole-benzimidazole derivatives was identified as a selective and potent pan-FGFR inhibitor.pyrazolo [4,3-b]pyrimido [4,5-e] [1,4]diazepine scaffold was identified as a selective and potent pan-FGFR and VEGFR-2 inhibitors(Fig.1c) [16][17][18] .
The biological and therapeutic properties of 5-amino-pyrazoles as well as their flexibility in synthetic processes have recently drawn a lot of attention and given them a prominent position 19 .The -NH 2 functionality makes 5-amino-pyrazoles one of a class of adaptable aromatic heterocycles containing nitrogen that have a high degree of controllability for creating different synthesis schemes.α-Aminophosphonates (AAPs) are most important group of organophosphorus compounds that are structural analogues of amino acids where a carboxylic substituent is substituted by phosphonic acid or related groups 20 .Recently, α-aminophosphonates have received enormous attention from scientific researchers in pharmaceutical and medicinal chemistry due to they have extremely high levels of wide biological activities such as anti-HIV, peptidemimics, antibacterial, antibiotics, inhibitors of serine hydrolase, herbicidal, antiviral, anticancer, enzyme inhibitors, antifungal, and anti-proliferative, enzyme inhibitors, inhibitors of serine hydrolase anti-Alzheimer and apoptosis inducing 21,22 .Moreover, α-aminophosphonate derivatives containing a pyrazole moiety showed a significant inhibitory effect on acetylcholinesterase (AChE) 23 .Vascular endothelial growth factor (VEGF) is a key signalling molecule that controls the tumor angiogenesis process.VEGF overexpression was discovered in a number of cancer, including Epdermoid carcinoma (HEP2) and Colon cancer (HCT-116).All VEGF responses in endothelial cells are mediated by VEGFR-2.Therefore, VEGFR-2 should be the primary target of any new drugs being developed to treat cancers in humans that are dependent on angiogenesis 24 .
Various methods for the synthesis of α-aminophosphonates were reported.However, one pot Kabachnik-Fields reaction, a one-pot multicomponent synthesis involving an amine, an aldehyde, and a phosphite in the presence of a Lewis acid catalyst remains the most efficient, simple, general, and high yielding method 25 .
Based on these facts and keeping in view the wide range of biological activities of pyrazole moieties, and aminophosphonate scaffolds, in this work, we expect that the incorporation of these moieties in the same scaffold structure may lead to good activities and potent powerful anticancer medicines.Thus, as a continuation of our prior work in the synthesis of biologically active heterocycles [26][27][28] , we have designed and synthesized a series of new α-aminophosphonates derivatives bearing pyrazole moiety and were evaluated against references and cancer cell lines.Additionally, the newly synthesized compounds' biological activity was examined in connection to changes in their molecular and electronic structure using density functional theory in an effort to connect theoretical and experimental results.To determine the target enzyme and the most active compound's mechanisms of action, a molecular docking simulation will be run.and 3).The structures of investigated α-aminophosphonates 2a-f and 4a-d, were confirmed by corrected elemental analysis, FT-IR, 1 H NMR and 13 C-NMR spectroscopy.The FT-IR spectrum was characterized by the following absorption bands: A band at 1267-1193 cm -1 is attributed to the stretching vibration of P=O group.υ (P-O-C) appeared at 1104-1011 cm -1 and υ (P-CH) is absorbed at 765-701 cm -1 .The CH aromatic stretching band is absorbed at 3162-3025 cm -1 , while the CH aliphatic bands are appeared at 2982-2853 cm -1 .Finally, NH/OH groups are absorbed at 3464-3413 cm -1 .

Synthesis of aminophosphonates and spectroscopic characterization.
The 1 H-NMR (DMSO) spectra of α-aminophosphonate derivatives, 2a-f, showed the following signals: δ 5.02-5.76 as singlet was assigned to P-CH and the singlet signals at the range of δ 6.23-7.12were attributed to pyrazole-CH.The Ar-H signals appeared at the range of δ 6.32-9.26ppm, also the -NH proton appeared at δ 5.61-5.99 and signals at δ 6.42-7.21were attributed to pyrazole-CH.The methylene and methyl protons of P-O-CH 2 CH 3 resonated as quarter and triplet respectively at δ 3.62-4.39and 1.26-1.39ppm.The 13   .Additionally, a peak at m/z corresponding to the product's molecular ion was visible in the mass spectrum for the all product.By using elemental microanalysis to characterize the examined substances, it was found that the calculated value and the measured value agreed well.The selected spectroscopic data are reported in the ("Experimental" section) (Figs.S1-S30 in supplementary materials).

Biological evaluation.
Antitumor activity.Nitrogen-containing heterocycles exhibit anticancer effect in various types of cancer through inhibiting cell growth and induction of cell differentiation and apoptosis.
Moreover, therapeutic drugs that inhibiting EGFR and VEGFR-2 can enhance the effectiveness of cancer therapy and resolve resistance issues 29 .
To create novel molecules with high inhibitory potentials, pyrazole derivatives were widely utilized 30 .The cytotoxic potency of the synthesized phosphonates, 2a-f and 4a-d was determined in vitro using the standard MTT method 31 against Colorectal carcinoma Colon cancer (HCT-116) and Epdermoid Carcinoma (HEP2) and also Human lung fibroblast normal cell line (WI38), using Doxorubicin as a positive control.The IC 50 values were estimated for each compound and the results are shown in (Figs. 4, 5 and 6) and summarized in (Table 1).Compounds 2a, 4b and 4d gave the highest activity for Epdermoid Carcinoma (HEP2), while the rest compounds have  moderate to weak activity against the same cell line as shown in Table 1.For colon carcinoma cells (HCT-116) Compounds 2a, 2d and 4b gave the strongest activity among all compounds.Finally, all the studied phosphonate compounds have moderate to weak cytotoxicity activity on the normal cell WI-38.
From the above data, we could conclude that compound 2a which contain 2, 3-diOH electron donating substituent and compounds 2d and 4b which have electron withdrawing NO 2 group, are the most potent derivatives against the tested cancer cell lines and also have low cytotoxic activity on the normal cell line (Table 1).
Structure-activity relationship (SAR) of studied compounds 2a-f and 4a-d.The experimental cytotoxicity of the investigated compounds to their structures was used to postulate a structure-activity relationship of the produced α-aminophosphonate derivatives: (i) The variety of substituents in the aryl aldehyde moiety of AAPs is important for the wide range of cytotoxic activity against different cell lines (HCT-116, Hep2 and WI-38), (Figs. 4, 5 and 6).(ii) Compound 2b which has one OH group in position 3 in the aryl aldehyde moiety is in more active compound towards the cell lines, but compound 2a which has 2-OH groups in position 2 and 3 in the aryl aldehyde moiety enhanced the cytotoxic activity and showed a medium cytotoxic activity against the normal lung cell WI-38 compared with doxorubicin which has OH moieties.(iii) Compound 2c which has one Cl group in position 3 in the aryl aldehyde moiety enhance the biological activity against the tested cell lines.(iv) The di-ethyl phosphonate, 2d, and the di-phenyl phosphonate compound 4b which have electron withdrawing substituent NO 2 in the aryl aldehyde moiety have potent cytotoxic activity against the cell line among all compounds.(v) The di-phenyl phosphonate compound 4d which has electron donating group N(CH 3 ) in the aryl substituted aldehyde moiety is more active than the di-ethyl phosphonate of the same substituent 2f which is may be due to the increasing number of phenyl rings which increase the resonance and accordingly increase the antitumor activity.(vi) The presence of two electron donating OH group in compounds 2a improves potency more than compound 2d which have an electron withdrawing group NO 2 (Table 1).

Molecular computational calculation. Geometry optimization and global reactivity descriptors using
DFT. in supplementary materials).
The "HOMO energy level" that occupies the highest molecular orbital is essentially an electron donor molecular orbital.The lowest unoccupied molecular orbital, or "LUMO", is mostly used as an electron acceptor.Both orbitals were called frontier molecular orbitals (FMOs) (Fig. 8 and Figs.S32 and S33 in supplementary materials) have evaluated kinetic stability, electronic transitions, and electro-optical properties 32 .The FMOs theory proposed that aromatic compounds have a coordination site (electrophilic attack).Furthermore, most reactions are caused by the interaction of one moiety's HOMO and another's LUMO.
Table 2 displays the energy gap (ΔE = E HOMO -E LUMO ), and chemical descriptors for the examined compounds.These descriptors can be assessed using the formulae provided by Pearson, and Padmanabhan et al. 33,34 .
The additional electronic charge is derived by the formula provided by Geerlings et al. and Pearson 33,34 .and is equal to the maximum number of electrons moved in the chemical process (ΔN max ).
Optimized molecular structure of 2e.
Using the computation findings provided in Table 2: a.The gas phase energy decreases with the order: 2b > 2a > 2d > 2f > 4a > 2e > 2c > 4b > 4c > 4d.This order indicates that the stability of 4d is higher than that of other compounds.b.The E HOMO and E LUMO values are both negative, suggesting that the isolated compounds are stable 35,36  that 2f has the greatest value (3.8900 e), which is greater than other compounds, showing its high electron acceptance.e.The dipole moment value of the 4a is greater than other compounds' values which may increase its hydrophilic nature and, as a result, its biological potency.

Computation of vibrational frequency.
A frequency calculation study was performed to determine the spectroscopic signature of substances (Fig. S34 in supplementary materials).Because the calculations were conducted for free molecules in a vacuum whereas the tests were performed for solid samples, there are modest discrepancies in theoretical and practical vibrational wavenumbers as displayed in (Fig. S34 in supplementary materials).Because of the poor symmetry of compounds, the modes of vibration are extremely complicated.Because of mixing with ring modes and substituent modes, in/out of plane, and torsion manners are the hardest to assign.Though, there are several strong frequencies in the IR spectrum that are important to define (Fig. S35 in supplementary materials) depicts the correlation graphic that described the agreement between theoretical and experimental wavenumbers.Table 3 shows the linear relationships between computed and experimental wavenumbers for all substances.For each figure, the correlation coefficient R 2 was obtained, where R 2 is a statistical number indicating how closely the regression line approximates the true data points.As a result, a good connection was discovered for the examined chemicals, as shown in (Fig. S35 in supplementary materials).Electrostatic potential V(r) and average local ionization energy I(r) of the investigated compounds have been demonstrated to be dynamic guides to its reactive performance 37 .Figure 9 and supplementary material (Figs.S36 and S37) depicted the electrostatic potential V(r) and average local ionization energy I(r) of all compounds.Also, computed molecular surface data recorded in Table 3 and the following parameters are listed in this table: a.The most positive V S,max and the most negative V S,min .b.The whole surface potential value V S , with its positive averages V + s and negative averages V − s .c.Most positive I S,max and most negative I S,min, and the average over the surface of the local ionization energy I S,ave.d.Internal charge transfer (local polarity) Π, is derived as a sign of internal charge separation and is present even in molecules with zero dipole moment due to symmetry.e.The variances,σ 2 + , σ 2 − and σ 2 tot describe the intensities and variations of the positive, negative, and overall surface potentials, respectively 38 .f.An electrostatic balancing parameter ν, that indicates the degree of equilibrium between positive and negative potentials; it has a maximum value of 0.25 when Table 4 shows that 2a has the largest Π with value 12.52 kcal/mol followed by 2b.This is because to their strong polar structures.While 2c exhibits lowest value of Π = 10.16 kcal/mol and may be related to its structural symmetry.
The positive surface potential ( V + s ) and negative surface potential ( V − s ) of 2a are strong with balanced with ν = 0.250.While, the weaker surface potentials with ν = 0.194 is related to 2e.Furthermore, the largest value of σ 2 tot for 2a shows the strong and variable + ve and -ve surface potentials.Figure 9, is presented the V S (r) and I S (r) on surfaces of 2a as a representative example and shows the loca- tions of the various most positive, V S,max and most negative, V S,min as well as the highest, I S,max and lowest, I S,min .On a particular molecular surface, there are frequently many local minima and maxima of each attribute.The highest negative electrostatic potential on the surface of 2a is related to the oxygen (O27), V S,min = − 39.1 kcal/ mol, followed by a weaker value of − 39.07 kcal/mol on the oxygen (O28).As a result, V S (r) incorrectly guess electrophilic attack to arise specially at the oxygen atom.However, the lowest value of I S (r) is found on the (C20), with I S,min = 189.27kcal/mol; as well as, there is an I S,min by the nitrogen (N31), but it is significantly higher, 198.28 kcal mol -1 .Thus, I S (r) demonstrates that the least-tightly-bound electrons, and most reactive are at (C20), correctly suggesting that these sites are more vulnerable to electrophiles.On contrast, the extremely significantly positive electrostatic potential of hydrogen (H50), V S,max = 61.78kcal/mol, and the V S,min = 7.16 kcal/mol of the hydrogen (H41) show their proclivity for noncovalent H-bonding as a donor.Molecular docking.Molecular docking with VEGFR-2 and FGFR1 proteins.The relevance of molecular docking in drug development is well acknowledged.According to the Glide S-score, the Maestro programme was used to dock the most extensive docking pocket and hits (Tables 5 and 6).Higher Glide S-score and lower RMSD values were used to identify the best docking hit compounds.Thus, the results in Table 5, reveals that the standard drug Doxorubicin exhibits the highest inhibitory activity against both proteins.Furthermore, the compounds 2b and 2a have the strongest inhibitory activity of the VEGFR2 and FGFR1 proteins, respectively.The sequence of inhibitory activity to the VEGFR2 protein is 2b > 2f > 4a > 2a > 4b > 4d > 2e > 2c > 2d.While the sequence of inhibitory action to the FGFR1 protein is 2a > 2b > 2f > 4d > 4a.Furthermore, 4c doesn't exhibit any interactions with the VEGFR-2 protein.Also compounds 4c, 2e, 2d, 4b, and 2c don't show any interactions with the FGFR1 protein.From these data, one can notice that compound 4c has no interaction with FGFR1 protein and agree with experimental data of in vitro cytotoxicity against WI38 (IC 50 > 100).The examined substances demonstrate that H-donor and H-acceptor interactions are the most common kind of interaction with VEGFR2 and FGFR1 receptors (Figs. 10, 11 and Figs.S38-S45 in supplementary materials).The interactions of chemical 2a with both proteins were shown as illustrative instances in (Figs. 10 and 11).From VEGFR-2-2a interaction (Fig. 10), hydrogen donor interaction from the hydroxyl group of 2a to water molecule with distance 1.83 Å, in addition, H-acceptor interactions from water molecule to (P-O) group, PTR1052 to (N=O) and LYS1053 to (N=O) with distances 1.97, 2.22 and 2.74 Å, respectively.While the interaction of 2a with FGFR1 protein (Fig. 11) exhibits hydrogen donor interaction from the hydroxyl group of 2a to ASP641 and hydrogen acceptor interaction from LYS514 to the same hydroxyl group of 2a with distances 1.63 and 2.39 Å, respectively.The good interactions of 2a with both receptors may be due to the highest positive and negative surface potentials of 2a obtained from electrostatic potential calculations.The molecular docking result for the interactions of investigated compounds, with both proteins reveals that the most common types of interactions are hydrophobic interactions, such as π-cation, as well as H-acceptor, H-donor, halogen bond, and salt bridge interactions.One possible source of the difference between the theoretical and experimental data is the limitations of molecular docking computational calculations.These calculations rely on mathematical functions, algorithms, and active site sensitivity to predict the interactions between drugs and biological targets.However, these calculations do not consider the factors that affect the drug delivery method or mechanism, the experimental variations, or the biological target properties 39 .and 2.92 Å as well as π-cation interaction between the aromatic ring and LYS514 with distance 5.51 Å.The re-docked original ligand was overlaid onto the native cocrystallized (Figs. 12 and 13).The re-docked inhibitor linked to VEGFR-2 and FGFR1 receptors has molecular docking scoring (-12.827 and -6.290 kcal mol −1 ) and the RMSD of (0.618 and 0.792 Å), respectively).

3D-QSAR studies.
To create the model, we employed the atom-based QSAR module.For all categories, we used four PLS factors.We used the cross-correlation coefficient 40 to assess the models' prediction abilities (Table 7).The comprehensive data for the QSAR investigation is shown in Table 8.Each set's best model result is presented in this table.
The models have R 2 , the regression coefficient, ~ 0.5-0.9, while the PLS factor 4 has R 2 ~ 0.9, indicating the model's robustness.The cross-correlation coefficient and regression coefficient are close, indicating that the model is stable.The models have a very high F variance and low P values, indicating that they are statistically significant.The cross-correlation coefficient Q 2 for the models is 0.635 and 0.724, respectively.This shows that the models are feasible.Figures 14 and 15 show the graphs for the models' training and test sets.
Blue cubes represent favorable locations toward inhibitory action in the models of VEGFR2 and FGFR1 receptors for substances (Fig. 16 and Figs.S42-S58, supplementary materials), whereas red cubes represent unfavorable regions.Compound 2a, for example, demonstrates the hydrogen bond donor action of NH and Table 5.Molecular docking glid G-scoring, RMSD, interaction results of the investigated compounds towards VEGFR-2 receptor (PDB ID: 1YWN).Glide G-score: Kcal/mol, RMSD and distance: Å.      www.nature.com/scientificreports/both phenolic OH groups for the VEGFR2 model but is not very effective for FGFR1 model suppression.The hydrogen bond donor effect of NH is not as good as VEGFR2 inhibition, and the hydrogen bond donor effect of one of the hydroxyl groups for the FGFR1 model is reduced while the impact of the other OH group is vanished.The compound's hydrophobic impact cannot be utilized to differentiate between VEGFR2 and FGFR1 inhibition.The majority of hydrophobic areas behave similarly to both systems.
The electron-drawing effect is likewise insufficient for distinguishing the inhibitory characteristics of VEGFR2 and FGFR1.Also, the electron-withdrawing effect of N(CH 3 ) 2 substituent is equally good for inhibiting both models.But the electron-withdrawing effect of (NO 2 ) substituents shows good and poor inhibiting.Finally, the (Cl) substituent doesn't exhibit any effect.The positive ionic effect of (NO 2 ) ion made compounds are not favorable for VEGFR2 inhibition.On the other hand, the positive ionic effect of (NO 2 ) toward FGFR1 inhibition is good inhibitor except for the compound 2f and 4d.As a result, the compound's positive ionic action on both inhibitory systems is very instructive in distinguishing between VEGFR2 and FGFR1 inhibition.

Conclusion
In conclusion, this study is based on synthesized and evaluated new series of α-aminophosphonates 2a-f and 4a-d derivatives containing pyrazole moiety by reaction of various aromatic aldehydes with diethyl and/or diphenyl phosphites in presence of LiClO 4 as a catalyst under convenient and efficient conditions via the Kabachnik-Fields reaction.The structure of the synthesized compounds was confirmed by elemental analysis, FT-IR, 1 H NMR, 13 C NMR, and MS spectral data.All of the synthesized compounds were evaluated for their in vitro antitumor activities, against colorectal carcinoma Colon cancer (HCT-116) and Epdermoid carcinoma (HEP2) and also Human lung fibroblast normal cell line (WI38).From the data, we could conclude that all these novel α-aminophosphonates could inhibit tumor cell lines (HCT-116 and HEP2) below 10 µM by the MTT assay.Moreover, Compounds 2a, 4b and 4d exhibited more potent inhibitory activity for Epdermoid Carcinoma (HEP2 with IC50 values (0.9 ± 12.5, 1.40 ± 17.76, and 1.8 ± 21.28 μM), respectively, compared with doxorubicin (0.6 ± 8.54 μM), while the rest compounds have moderate to weak activity against the same cell line.For colon carcinoma cells (HCT-116) Compounds 2a, 2d and 4b gave the strongest activity among all compounds with IC50 values (0.7 ± 9.51, 1.6 ± 19.90, and 1.1 ± 15.06 μM), respectively, compared with doxorubicin (0.3 ± 5.32 μM).Finally, all the studied phosphonate compounds have good to moderate cytotoxicity activity on the normal cell WI-38.The geometry optimization results using DFT exhibits that the stability of 4d is higher than that of other compounds as well as 2f has the highest electron acceptance while dipole moment of 4a is greater than other compounds' values which may increase hydrophilic nature of 4a.From molecular docking interaction, the results reveals that compounds 2a shows good interactions with VEGFR-2 and FGFR1 proteins respectively, this behavior may be due to the highest positive and negative surface potentials of 2a obtained from electrostatic potential calculations.On the other hand, compound 4c doesn't exhibit any interactions with the VEGFR-2 protein.Also compounds 4c, 2e, 2d, 4b, and 2c don't show any interactions with the FGFR1 protein.From these data, one can notice that compound 4c has no interaction with FGFR1 protein and agree with experimental data of in vitro cytotoxicity against WI38 with IC50 values (> 100 μM).

Experimental
Materials and instrumentation.All data of chemicals and instruments are available in the supplementary file (Sect.S1).In accordance with literature procedures, the compounds aryl substituted pyrazolamines 41 were synthesized.

Figure 7 and
Fig. S31 in supplementary materials illustrate the molecular structure and atom numbering of the examined compounds.One can come to the conclusion that the values of bond lengths and angles are close to the actual values based on the analysis of the data generated for bond lengths and angles (Tables S1-S20 . c. Hard molecules have larger HOMO-LUMO disparities, whereas soft and active molecules have smaller energy differences.The chemical potential 'μ' which quantifies electrons' ability to escape from the equilibrium structure, is decreased as follows: 4a (− 3.835 eV) > 2a (− 3.835 eV) > 2b (− 3.925 eV) > 2c (− 4.021 eV) > 2e (− 4.162 eV) > 4d (− 4.167 eV) > 2f (− 4.214 eV) > 2d (-4.378 eV) > 4b (− 4.445 eV) > 4c (− 4.482 eV).d.ΔN max is the reaction index, which gauges the bond energy's stability.The computation of ΔN max reveals

Figure 9 .
Figure 9.The surface structure of ESP and ALIE using DFT method for 2a.

Figure 14 .
Figure 14.Experimental vs predicted activity of atom-based QSAR model for VEGFR2 model.

Figure 15 .
Figure 15.Experimental vs predicted activity of atom-based QSAR model for VEGFR2 model.

Figure 16 .
Figure 16.QSAR models of compound 2a with VEGFR2 and FGFR1 receptors inhibition for (a) hydrogen bond donor effect, (b) hydrophobic effect, (c) positive ionic effect, and (d) electron withdrawing effect.

Table 2 .
Calculated gas phase energy, and energetic descriptors for investigated compound.

Table 3 .
Comparison of experimental and theoretical IR spectra of investigated compounds.

Table 7 .
Prediction of QSAR models on drug compounds.

Table 8 .
Statistical parameters of atom-based QSAR model.