INTRODUCTION:

c-SRC tyrosine kinase is a member of SRC family kinases (SFKs) which plays a crucial role in cell signaling and regulation of cellular processes¹. It is involved in various biological functions, including proliferation, migration and apoptosis.

c-SRC kinase regulates cell growth and proliferation by modulating signaling pathways involved in cell cycle progression². It interacts with key regulators of cell cycle checkpoints and promotes cell division. It phosphorylates and activates focal adhesion kinase (FAK) and other proteins involved in cytoskeletal remodeling, leading to loss of adhesion and cell migration.

c-SRC kinase has been shown to regulate cell survival and apoptosis pathways by influence in anti-apoptotic and pro-apoptotic signaling cascades³. Moreover, c-SRC kinase plays a role in angiogenesis by involving in signaling pathways that regulate endothelial cell migration and tube formation.

Dysregulation of c-SRC tyrosine kinase involves in various diseases, particularly cancer. Aberrant activation or overexpression of c-SRC kinase can promote tumor growth, metastasis, and drug resistance⁴. It is involved in multiple signaling pathways that contribute to cancer progression, including those related to cell proliferation, survival, angiogenesis, and invasion⁵.

Traditional drug design process can be very expensive, while new technologies known as Computer Aided Drug Design (CADD) are less time and cost consuming with more efficiency⁶. CADD are used in different disciplines such as molecular biology, biochemistry, etc^7-9. CADD is applicable in cancer disease treatment and it is considered as an effective tool in search and development of new drugs^10,11.

CADD is classified into two types: Structure Based Drug Design (SBDD) and Ligand Based Drug Design (LBDD)¹². Structure Based Drug Design depends on 3D structure of the target such as enzymes, proteins or receptors, we can visualize the binding mode of ligands with the target, predicting the locations of key binding pockets and the affinity of the ligand to the target of biological interest¹⁰.

While in LBDD the crystal structure of the target is unknown¹⁰.Pharmacophore modeling, molecular similarity approaches, and QSAR (quantitative structure–activity relationship) modeling are the most common methods of this type¹³.

Quantitative Structure-Activity Relationship (QSAR) represents a powerful computational approach that correlates the physicochemical properties of molecules with their biological activities¹⁴. By establishing a mathematical relationship between the structural features of a compound and its biological activity, QSAR models can predict the inhibitory potency of new compounds against the targeted enzyme¹⁵. This enables the identification of lead compounds with improved activity and selectivity, thereby guiding the rational design of novel inhibitors¹⁶.

The application of QSAR in drug discovery has gained significant attention due to its ability to guide the design of structurally diverse compounds with desired properties. By utilizing QSAR models, researchers can prioritize compounds for synthesis, optimize lead compounds, and explore potential modifications to improve potency, selectivity, and other desirable drug-like properties¹⁶.

Different drugs currently have been discovered or developed based on ligand-based methods such as Zolmitriptan, Norfloxacin, and Losartan. Norfloxacin is a drug used in urinary tract infections, it was developed using QSAR model and was approved by Food and Drug Administration in 1986, while Losartan is used as hypertension and Zolmitriptan is used to treat migraines¹⁷.

Researchers can utilize various computational techniques to generate QSAR models, such as multiple linear regression (MLR), partial least squares (PLS), and machine learning algorithms⁸ like random forest, support vector machines, or neural networks¹⁸. The choice of the modeling approach depends on the complexity of the dataset and the specific research objectives.

In this study, we aimed to use QSAR modeling techniques to establish a new QSAR equation between the physicochemical properties of molecules with their biological activities. We used different inhibitors with known biological activity toward c-Src to construct robust QSAR model. The resulting model can be used to predict the inhibitory potency of new compounds and guide the rational design of novel c-src inhibitors with improved activity and selectivity.

METHODS:

Dataset selection: A diverse dataset of known c-SRC tyrosine kinase inhibitors and their biological activities is compiled from literature sources and/or proprietary databases^19-21.

Molecular descriptors calculation: A set of molecular descriptors is calculated for different R₂ substitution and the entire compound in the dataset using Accelrys Discovery Studio 3.5 software was used to capture the physicochemical, structural, and electronic properties of the compounds (Table 2 and 3).

QSAR Equation Development:

To select out the predominant descriptors affecting the Src inhibitory activity of the compound, the correlation analysis was performed by the statistical software SPSS by using each candidate as an independent variable and pIC₅₀ as a dependent variable. The descriptors with higher correlation with the pIC₅₀ were selected to carry out the stepwise multiple linear regression analysis to establish the optimal 2D-QSAR model. Then, the well-known scheme of leave-one-out (LOO) cross-validation was adopted to evaluate the predictive ability of the established equation. This test is necessary because a high correlation coefficient R2 only indicates how well the equation fits the data rather than how well it can predict any new data not included in the fitted data. The square of cross-validation coefficient q², which is used as a criterion of both robustness and predictive ability of the model, should be >0.5 for a reliable model²². To obtain QSAR with more reliable predictive ability, external validation is also crucial, and it was evaluated by the test set to farther confirm its predictive ability. A QSAR model is accepted to own high predictive power only if the square of predictive correlation coefficient (R² pred) between the experimental and predicted activities is >0.6 for the test set²².

RESULTS AND DISCUSSION:

Dataset collection:

A set of 22 variously functionalized pyrrolo-pyrimidine c-Src inhibitors were collected from the literature along with their activity data (Table 1). The IC₅₀ values were converted to pIC₅₀ (-log IC₅₀) values, and used as dependent variables in the QSAR analysis. Then, compounds with R₁ substitution (R₁=m-OCH3) were selected in order to study the effect of R₂ substitution on efficacy.

Figure 1: Pyrrolo-pyrimidine c-Src inhibitors general structure.

The compounds were divided into two groups: Training set and Test set. Training set compounds were selected following the usual guidelines (compounds belonging to the training set should be representative of the molecular diversity of all the compounds under study and uniformly span over the whole range of activity).

Table 1: The compounds used in the study and their pIC₅₀.

#	R1	R2	pIC50
1	H	H	7
2**	m-OCH₃	p-CH₂CH₂OH	7
3*	m-OCH₃	p-CH₂CH₂(4-OH piperridin-1-yl)	7.69897
4	m-OCH₃	H	7.30103
5	p-COOCH₂CH₃	H	5
6*	H	o-CH₂OH	6.52288
7	H	m-CH₂OH	6.22185
8	H	p-CH₂OH	7.69897
9	m-OCH₃	m-CH₂OH	7.22185
10	H	o-CH₂NHCH₂CH₂OH	5.30103
11	H	m-CH₂NHCH₂CH₂OH	8
12	H	p-CH₂NHCH₂CH₂OH	7.52288
13**	m-OCH₃	p-CH₂NHCH₂CH₂OH	7.52288
14*	m-OCH₃	p-CH₂CONHCH₂CH₂OH	7.69897
15*	m-OCH₃	p-CH₂CH₂N(CH₃)CH₂CH₂OH	7.52288
16*	m-OCH₃	p-CH₂COOH	7.30103
17*	m-OCH₃	p-CH₂CH₂N(CH₃)₂	7.69897
18**	m-OCH₃	p-CH₂CH₂NHCH₂CH₂OH	8
19*	m-OCH₃	p-CH²CH²N(CH³)CH²CH²OCH³	8
20**	m-OCH₃	p-OCH₂CH₂NHCH₂CH₂OH	7.69897
21*	m-OCH₃	p-OCH₂CH₂NHCH₂CONHCH₃	7.39794
22	m-OCH₃	p-OCH₂CH₂ imidazol-1-yl	6.82391

*Training set

**Test set

Descriptors Calculation:

Feature importance analysis was performed to identify the descriptors that significantly contributed to the inhibitory activity against c-SRC tyrosine kinase. Among the descriptors, Jurs_PPSA_1 (R₂), Jx(R₂) and HBA_Count (Whole molecule) were found to be the most influential features.

Table 2: Statistical results of the correlation between pIC50 and the best descriptors for the substituent R₂.

		Molecular_Weight	Jurs_PPSA_1	Molecular_SASA	Molecular_SAVol	Molecular_SurfaceArea	Molecular_Volume	Num_H_Acceptor	Jx	Jy
pIC50	Pearson correlation	0.778^**	0.752^*	0.692^*	0.692^*	0.780^**	0.796^**	0.667^*	0.699^*	0.668^*
	Sig. (2-tailed)	0.008	0.012	0.027	0.027	0.008	0.006	0.035	0.036187	0.049
	N	8	8	8	8	8	8	8	8	8

*. Correlation is significant at the 0.05 level (2-tailed).

**. Correlation is significant at the 0.01 level (2-tailed).

Table 3: statistical results of the correlation between pIC₅₀ and the best descriptors for the entire compound.

		Molecular_3D_SAVOL	Molecular_SAVol	Molecular_SurfaceArea	Molecular_Volume	Molecular_Weight	HBA_Count	Molecular_3D_SASA	Molecular_Mass	Jx
pIC50	Pearson correlation	0.805^**	0.793^**	0.800^**	0.808^**	0.784^**	0.646^*	0.803^**	0.784^**	0.768^**
	Sig. (2-tailed)	0.005	0.006	0.005	0.005	0.007	0.044	0.005	0.007	0.009
	N	8	8	8	8	8	8	8	8	8

*. Correlation is significant at the 0.05 level (2-tailed).

**. Correlation is significant at the 0.01 level (2-tailed).

The obtained QSAR model demonstrates the significance of Balaban index for substituent R₂. The descriptor Balaban index is a type of topological index that represents extended connectivity, it is a good descriptor for the shape of the molecules. The direct relationship between Balaban index of R₂ substituent and inhibitory potencyindicates that a bigger size and high branching R₂ substituent increase the inhibition activity.

Similarly, the positive correlation between the surface of partial positively charged surface area (Jurs_PPSA_1) and inhibition activity suggests that more positive compounds tend to exhibit higher inhibitory potency. This could be attributed to increased binding interactions and improved target engagement. Additionally, the number of hydrogen bond acceptors was positively correlated with activity, indicating the importance of hydrogen bonding in the interaction between the inhibitors and the target enzyme.

QSAR equation has been established depending on the descriptors that significantly contributed to the inhibitory activity against c-SRC tyrosine kinase.

pIC₅₀= 0.4802 (J_X)+ 0.003508 (Jurs_PPSA_1) + 0.1073(HBA_count) + 5.91

The model showed a good correlation coefficient (R² = 0.824) as well as a satisfactory internal and external predictive powers as shown in graph 1 and 2 respectively. The internal validation procedure (a cross-validation routine with 5 random sets of compounds) resulted in a cross-validated correlation coefficient (q²) of 0.609. all residuals fell within a statistically tolerable error range as shown it Table 4 (RMS residual error= 0.17021)²².

Table 4: The residuals values between the experimental and predicted pIC50 for the trainning set.

Inhibitor	Exp IC50	Pred IC50	Residuals
3	7.69897	7.58117	0.117804
9	6.52288	6.56293	-0.0400505
17	7.69897	7.45596	0.243008
18	7.52288	7.72679	-0.2039.9
19	7.30103	7.24752	0.0535088
20	7.69897	7.52525	-0.127309
22	8	7.80305	0.196955
31	7.39794	7.63795	-0.240007

Graph 1: Experimental activity versus predicted activity in the final QSAR model.

To ensure QSAR with more reliable predictive ability, external validation using test set compound (Table 1) has been established.

Table 5: Validation results of the QSAR model using external test set.

Validation Result Using External Test Set
Model Name	q²	RMS Error	Mean Absolute Error
pIC50	0.982	0.216	0.207

Graph 2: Experimental activity versus predicted activity for the external test set.

Based on the findings of Structure-Activity Relationship (QSAR) analysis, several new compounds have been proposed as c-SRC enzyme inhibitors. The substitutions at the R₂ position were diversified to align with the QSAR equation, particularly considering branching (J_x) and partial positive surface area (Jurs_PPSA1). Additionally, functional groups were introduced to act as hydrogen bond acceptors in order to maximize inhibition activity. Consequently, the following substituents were proposed:

Figure 2: General structure of proposed compounds as potential c-SRC tyrosine kinase inhibitors.

Table 6: Proposed compounds as potential c-SRC tyrosine kinase inhibitors.

#	R₂ substituent	#	R₂ substituent
1		10
2		11
3		12
4		13
5		14
6		15
7		16
8		17
9		18

As expected, the calculation of molecular descriptors of the compounds and their application to the QSAR equation revealed an increase in the predicted activity values (pIC₅₀) compared to the compounds originally used to derive the equation. The table below displays the predicted activity values for the newly proposed compounds:

Table 7: proposed compounds and their inhibition activity.

Compound	Predicted pIC50	Compound	Predicted pIC50
1	8.421	10	8.163
2	8.268	11	8.21
3	8.48	12	8.11
4	8.483	13	8.044
5	8.489	14	8.028
6	8.24	15	8.432
7	8.226	16	7.886
8	8.217	17	8.166
9	8.109	18	8.669

Based on the previous results, it is clear that compounds 1, 3, 4, 5 15 and 18 exhibits the highest potency. compound 1 is characterized by a high partial positive surface area due to the presence of the piperazine ring (the same applies to compound 8, Figure 3b). However, compound 1 includes a long linker with an oxygen atom, which enables it to achieve additional hydrogen bonds with amino acid residues Asp348 and Lys343, in addition to the primary bonds with Lys295, Glu339 and Met341 (Figure 3a).

Figure 3: a) Compound 1 interactions with active site residues, b) Compound 8 interactions with active site residues.

Compounds 4 and 5 exhibit significant branching, where the diethyl amine group contributes both significant steric bulk and a positive charge surface. Similarly, compound 15 also shows high branching due to the amide and phenol groups and a partial positive surface due to the same previously mentioned groups.

As for compound 18, it combines the three properties together due to the presence of an oxygen atom in the linker and an urea group that provides the partial positive charged surface and suitable branching, this link can achieve several bonds with the active site which explain its higher effectiveness compared to other compounds (Figure 4).

Figure 4: Compound 18 orientation in the active site.

CONCLUSION:

This study successfully identified key molecular descriptors influencing the inhibitory activity of pyrrolo-pyrimidine c-Src inhibitors. Among the most significant descriptors were Jurs_PPSA_1, Num_H_Acceptors, and HBA_Count, which highlight the critical roles of positively charged solvent-accessible surface areas and hydrogen bond acceptors in enhancing inhibitor potency. The established QSAR model, demonstrating a strong correlation coefficient (R² = 0.824) and satisfactory internal (q² = 0.609) and external validation metrics, underscores the reliability and predictive power of the model. These findings are consistent with previous research and provide valuable insights for the design and development of more effective c-Src tyrosine kinase inhibitor.

REFERENCES:

1. Thomas SM, Brugge JS. Cellular functions regulated by Src family kinases. Annu Rev Cell Dev Biol. 1997; 13: 513-609 doi:10.1146/annurev.cellbio.13.1.513.

2. Frame MC. Src in cancer: deregulation and consequences for cell behaviour. Biochim Biophys Acta. 2002; 1602(2): 114-30 doi:10.1016/s0304-419x(02)00040-9.

3. Kim LC, Song L, Haura EB. Src kinases as therapeutic targets for cancer. Nat Rev Clin Oncol. 2009; 6(10): 587-95 doi:10.1038/nrclinonc.2009.129.

4. Yeatman TJ. A renaissance for SRC. Nat Rev Cancer. 2004; 4(6): 470-80 doi:10.1038/nrc1366.

5. Summy JM, Gallick GE. Src family kinases in tumor progression and metastasis. Cancer Metastasis Rev. 2003; 22(4): 337-58 doi:10.1023/a:1023772912750.

6. Dubey R, Chandraker G, Sahu P, Paroha S, Sahu D, Verma S, et al. Computer aided drug design: A review. 2011; 2(3): 104-8

7. Bairagi A, Singhai AK, Jain AJAJPT. Artificial Intelligence: Future Aspects in the Pharmaceutical Industry an Overview. 2024; 14: 237-46

8. Kalkotwar R, Saudagar RJAJoPR. Design, Synthesis and anti microbial, anti-inflammatory, Antitubercular activities of some 2, 4, 5-trisubstituted imidazole derivatives. 2013; 3(4): 159-65

9. Dighe NS, Shinde P, Anap H, Bhawar S, Musmade DSJAJoPR. QSAR Study and Synthesis of some new 2, 5-disubstituted 1, 3, 4-oxadiazole derivatives as Anti-microbial and Anti-inflammatory Agents. 2014; 4(4): 174-9

10. Lionta E, Spyrou G, Vassilatis DK, Cournia Z. Structure-based virtual screening for drug discovery: principles, applications and recent advances. Curr Top Med Chem. 2014; 14(16): 1923-38 doi:10.2174/1568026614666140929124445.

11. Ganatra S, Patle M, Bhagat GJAJoRiC. Studies of Quantitative Structure-Activity Relationship (QSAR) of Hydantoin Based Active Anti-Cancer Drugs. 2011; 4(10)

12. Prema V, Sivaramakrishnan M, Rabiya M. A Concise Review on role of QSAR in Drug Design. 2023

13. Dighe AS, Tajamulhaq AEJAJoPR. An Overview of Molecular Docking. 2024; 14(3)

14. Sapkale G, Khandare D, Patil S, Surwase USJAJoRiC. Drug Design: An Emerging Era of Modern Pharmaceutical Medicines. 2010; 3(2): 261-4

15. Bhardwaj S, Dubey S. Qsar and Docking Studies of Some Novel Piperine Analogues as Monoamine Oxidase Inhibitors. 2022

16. Cherkasov A, Muratov EN, Fourches D, Varnek A, Baskin, II, Cronin M, et al. QSAR modeling: where have you been? Where are you going to? J Med Chem. 2014; 57(12): 4977-5010 doi:10.1021/jm4004285.

17. Hassan BM, Kumar GG. In silico chemistry and biology: current and future prospects/edited by Girish Kumar Gupta and Mohammad Hassan Baig.

18. Prajapati PM, Shah YR, Sen DJJRJoS, Technology. Artificial Neural Network: A New Approach for QSAR Study. 2011; 3(1): 17-24

19. Altmann E, Widler L, Missbach M. N(7)-substituted-5-aryl-pyrrolo[2,3-d]pyrimidines represent a versatile class of potent inhibitors of the tyrosine kinase c-Src. Mini Rev Med Chem. 2002; 2(3): 201-8 doi:10.2174/1389557023406188.

20. Missbach M, Altmann E, Widler L, Šuša M, Buchdunger E, Mett H, et al. Substituted 5, 7-diphenyl-pyrrolo [2, 3d] pyrimidines: potent inhibitors of the tyrosine kinase c-Src. 2000; 10(9): 945-9

21. Missbach M, Jeschke M, Feyen J, Muller K, Glatt M, Green J, et al. A novel inhibitor of the tyrosine kinase Src suppresses phosphorylation of its major cellular substrates and reduces bone resorption in vitro and in rodent models in vivo. Bone. 1999; 24(5): 437-49 doi:10.1016/s8756-3282(99)00020-4.

22. Tropsha A, Gramatica P, Gombar VKJQ, Science C. The importance of being earnest: validation is the absolute essential for successful application and interpretation of QSPR models. 2003; 22(1): 69-77

Received on 15.04.2025 Revised on 09.08.2025

Accepted on 06.10.2025 Published on 20.05.2026

Available online from May 25, 2026

Research J. Pharmacy and Technology. 2026;19(5):2059-2064.

DOI: 10.52711/0974-360X.2026.00295

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Creative Commons License.