Optimization of biological activities of Agaricus species: an artificial intelligence-assisted approach

Abstract

This study aims to determine the optimum extraction conditions that maximize the biological activities of Agaricus campestris and Agaricus bisporus species. In the study, a total of 64 extraction experiments were carried out at different temperatures, time and solvent concentrations and the obtained data were modeled with Artificial Neural Network (ANN) and optimized with Genetic Algorithm (GA). After determining the optimum extraction parameters, antioxidant, anticholinesterase and antiproliferative activities and phenolic contents of the produced extracts were analyzed. As a result of single and multi-objective optimization studies of A. campestris and A. bisporus, extract concentration was determined. In antioxidant analyses, it was observed that A. campestris extracts had higher total antioxidant capacity (TAS) and lower total oxidant level (TOS) compared to A. bisporus extracts. Anticholinesterase activity tests revealed that A. campestris extracts showed stronger inhibitory effect compared to A. bisporus. In addition, in the antiproliferative activity analyses performed on the A549 lung cancer cell line, it was determined that the extracts produced with ANN-GA optimization suppressed cell proliferation. Phenolic compound analyses showed that antioxidant compounds such as gallic acid, protocatechuic acid and caffeic acid were found at high levels in A. campestris extracts. The results show that ANN-GA supported optimization processes enrich the bioactive components of mushroom extracts and that these methods may be effective in biotechnological applications.

Introduction

Mushrooms are widely distributed heterotrophic organisms that play essential ecological roles. As decomposers, they contribute to the nutrient cycle by breaking down organic matter, thereby enhancing soil fertility. Beyond their ecological function, certain mushroom species are recognized for their significant medicinal and nutritional value, whereas others contain toxic components¹. Many mushrooms are rich in bioactive compounds exhibiting antioxidant, antimicrobial, and anticancer properties. These attributes make them highly valuable for applications in the pharmaceutical, food, and cosmetic industries^2,3. Accordingly, mushrooms are considered important natural resources with high potential in biotechnology, medicine, and agriculture.

Optimizing the extraction of bioactive compounds from mushrooms has become a critical area of research. The efficiency of extraction is influenced by several variables, including mushroom species, solvent type, temperature, duration, and pH^4,5. Accurate determination of these conditions ensures the effective isolation of biologically active compounds. ptimization strategies are often designed for single or multiple biological targets. Multi-objective optimization enables the enhancement of several biological properties—such as antioxidant, antimicrobial, or anticancer activities—simultaneously⁶. As a result, optimized mushroom extracts can yield more potent and commercially promising products. Although previous studies have mainly focused on optimizing antioxidant properties using artificial intelligence (AI)-based models, the present study expands this framework by incorporating antiproliferative and anticholinesterase evaluations. This integrated approach enhances the biomedical relevance and application potential of the extracts.

The aim of this study was to determine the optimal extraction conditions to maximize the biological activities of Agaricus campestris and Agaricus bisporus. Under the optimized conditions, antiproliferative, anticholinesterase, antioxidant activities, and phenolic compound contents of the extracts were comprehensively analyzed.

Material and method

The mushroom samples used in the study were collected from Antalya/Turkey. After the identification procedures, the most suitable extraction conditions were determined to show the highest biological activity of the mushroom samples. The fungarium samples of the mushrooms are kept in Osmaniye Korkut Ata University, Department of Biology. The identification of the samples was made by Dr. Mustafa SEVİNDİK.

Extraction procedure methods

Using a Soxhlet apparatus, 64 experiments were performed under different conditions, including extraction temperatures (40, 50, 60, and 70˚C), extraction durations (3, 5, 7, and 9 h), and extraction concentrations (0.5, 1.0, 1.5, and 2 mg/mL). The collected data was then modeled with an Artificial Neural Network (ANN) and optimized through a Genetic Algorithm (GA).

ANN and GA-based modeling approach

Modeling was carried out through the ANN approach. The model’s inputs included extraction temperature, extraction time, and extract concentration, with the TAS and TOS values of extracts serving as the output. Prior to modeling, all input and output variables were normalized using min–max scaling to the [0,1] range. This normalization procedure ensured that all variables contributed proportionally to the learning process and enhanced the convergence performance of the ANN and GA models.The structure of the ANN layers used in this work is illustrated in Fig. 1.

Fig. 1.

Layers of ANN model of study.

The data derived from experimental work was divided as follows: 80% for training, 10% for validation, and 10% for testing. The learning process utilized the Levenberg-Marquardt (LM) algorithm. The Levenberg–Marquardt (LM) algorithm was selected due to its fast convergence and high accuracy, especially in small-to-medium sized datasets typical in biological optimization studies. A momentum coefficient of 0.5 was employed to stabilize learning by reducing oscillations and ensuring smoother convergence. These values were based on preliminary trials and are supported by previous studies utilizing similar data structures and biological modeling approaches. A comparison of 20 different hidden neuron numbers (from 1 to 20) was conducted to find the optimal network. The learning and momentum coefficients were both fixed at 0.5, the maximum number of iterations was set to 500, the validation check limit was 50, and the error threshold was established at 1 × 10⁻⁵. During training, there is a risk of the performance surface becoming stuck at a local minimum, meaning a single training session might not achieve the best results. Thus, to approximate a global minimum, the networks were repeatedly retrained several times to secure the most effective network. For this research, each model was trained a total of 1000 times.

The study evaluated the effectiveness of the created models by employing mean square error (MSE) and mean absolute percentage error (MAPE) as performance measures. These were determined using the formulas provided in Eq. 1 (1) and Eq. 2 (2):

1

2

where is the experimental result, is the prediction result, and is the number of samples.

Optimization using ANN-GA and RSM

The optimization studies were conducted using the MATLAB program. Experiments were performed with varying population sizes, utilizing the roulette wheel technique for natural selection and the single-point crossover method for crossover operations. Convergence graphs were analyzed to determine the appropriate number of iterations. Each optimization process was repeated at least 60 times to achieve results very close to the global optimum. For the NSGA-II based multi-objective optimization, the following parameters were used: population size of 100, 200 generations, a crossover probability of 0.8, and a mutation probability of 0.05. The roulette wheel method was used for selection, and single-point crossover was employed. No additional constraint-handling strategy was needed, as the search space was bounded within biologically feasible limits based on prior experimental results.

In this study, both single-objective and multi-objective optimization analyses were conducted for two mushroom species. The single-objective optimization focused solely on maximizing the TAS value. In contrast, the multi-objective optimization aimed to maximize the TAS value while simultaneously minimizing the TOS value. For this reason, the multi-objective version of the genetic algorithm, known as the Non-Dominated Sorting Genetic Algorithm II (NSGA-II), was employed⁷. The final single solution from the set of solutions obtained in the multi-objective optimization was determined using the LINMAP method. The LINMAP (Linear Programming Technique for Multidimensional Analysis of Preference) method was used to select the final solution from the Pareto front. In this study, equal importance was assigned to both objective functions (TAS and TOS), and the Euclidean distance from the ideal point (maximum TAS, minimum TOS) was minimized. No additional utility function was applied beyond this standard linearization.

Extraction processes

Optimization studies were carried out to determine the extraction conditions under which mushroom samples would exhibit the highest biological activity and the most suitable extraction parameters were determined. As a result of the single-objective optimization for Agaricus campestris, the optimum extraction conditions were determined as 50.002 °C temperature, 4.615 h of time and 1.952 mg/mL extract concentration. Within the scope of the multi-objective optimization, the temperature was determined as 51.591 °C, 6.141 h of time and 1.769 mg/mL extract concentration. As a result of the single-objective optimization study for Agaricus bisporus, the most suitable extraction conditions were determined as 49.459 °C temperature, 4.983 h of time and 1.960 mg/mL extract concentration. In the multi-objective optimization study, the optimum values ​​were determined as 54.359 °C temperature, 7.355 h of time and 1.683 mg/mL extract concentration. The extraction process was carried out under conditions closest to the determined optimum extraction conditions. In this process, computer-aided optimum extraction settings were made using the Gerhardt SOX-414 device and extracts were obtained. Biological activity tests were performed on the extracts obtained under these conditions. After the extraction process, ethanol used as a solvent was removed with the Buchi R100 Rotary Evaporator and crude extracts were obtained.

Antiproliferative activity test

The antiproliferative effect of mushroom extracts obtained under optimum conditions was evaluated against A549 lung cancer cell line. In the study, solutions were prepared from the extracts at concentrations of 25, 50, 100 and 200 µg/mL. When the cells reached 70–80% confluence, they were dissociated using 3.0 mL Trypsin-EDTA solution (Sigma-Aldrich, MO, USA) and seeded on culture plates. Then, they were incubated for 24 h. After the incubation period, previously prepared stock solutions were added to the cells and the incubation was performed again for 24 h. Afterwards, the supernatants were removed, and the growth medium was replaced with 1 mg/mL MTT (Sigma) and incubated at 37 °C until purple precipitate formed. After completion of the reaction, the MTT solution was dissolved with dimethyl sulfoxide (DMSO) (Sigma-Aldrich, MO, USA) and the absorbance values ​​were measured at 570 nm using an Epoch spectrophotometer (BioTek Instruments, Winooska, VT)⁸.

Anticholinesterase activity test

The anticholinesterase activity of mushroom extracts obtained under optimum conditions was analyzed using the Ellman method⁹. Galantamine was preferred as the standard compound. Stock solutions were prepared from mushroom extracts in the concentration range of 3.125-200 µg/mL. In the experimental phase, 130 µL of 0.1 M pH 8 phosphate buffer, 10 µL of stock solution and 20 µL of enzyme solution (AChE or BChE) were added to the microplates, respectively, and incubation was carried out for 10 minutes at 25°C in the dark. After incubation, 20 µL of DTNB (5,5’-dithiobis-(2-nitrobenzoic acid)) solution and 20 µL of substrate (acetylcholine iodide or butyrylcholine iodide) were added and absorbance measurements were made at 412 nm. All samples were analyzed in triplicate. Inhibition percentages of the samples were calculated based on the data obtained and IC50 values ​​were expressed in µg/mL.

Antioxidant activity tests

Total antioxidant and oxidant analysis

The total antioxidant capacity (TAS) and total oxidant level (TOS) of mushroom extracts obtained under optimum conditions were analyzed using Rel Assay kits. The tests were performed in accordance with the protocol provided by the kit. TAS results were reported as mmol Trolox equivalent/L, while TOS values ​​were expressed as µmol hydrogen peroxide equivalent/L^10,11. Oxidative Stress Index (OSI) was determined by calculating the percentage of TOS and TAS values¹².

DPPH free radical scavenging activity

Mushroom extracts obtained under optimum conditions were dissolved in DMSO to form stock solutions with a concentration of 1 mg/mL. 1 mL of these stock solutions were mixed with 160 µL of 0.267 mM DPPH solution prepared in 4 mL of 0.004% methanol. The mixture was left at room temperature in a dark environment for 30 min. At the end of the period, the absorbance of the reaction mixture was measured at a wavelength of 517 nm. The antioxidant capacity of the obtained extracts was calculated in mg Trolox equivalent/g¹³.

Ferric reducing antioxidant power assay

Optimized solutions obtained from mushroom extracts were prepared as stock solutions in a volume of 100 µL. A certain amount of this stock solution was mixed with 2 mL of FRAP reagent. FRAP reagent was prepared by mixing 300 mM acetate buffer (pH 3.6), 40 mM HCl, 20 mM FeCl₃ 6 H₂O solution and 10 mM 2,4,6-tris(2-pyridyl)-S-triazine solution at a ratio of 10:1:1. The prepared mixture was incubated at 37 °C for 4 min. Then, absorbance measurement was performed at a wavelength of 593 nm and the results were expressed as mg Trolox equivalent/g¹⁴.

Phenolic analysis

The phenolic compound profile of the optimized mushroom extracts was analyzed by LC-MS/MS method. In this context, 24 different standard compounds found in the extracts were evaluated. The separation of the compounds was carried out using C-18 Intersil ODS-4 (3.0 mm × 100 mm, 2 μm) analytical column operating at 40 °C. Water containing 0.1% formic acid (Phase A) and methanol containing 0.1% formic acid (Phase B) were used as the mobile phase. The flow rate was determined as 0.3 mL/min during the analysis, and the injection volume was set as 2 µL.

Statistical evaluation

Statistical analyses were performed using SPSS 21.0 for Windows. One-way analysis of variance (ANOVA) followed by Duncan’s multiple range test was used to determine significant differences among extract groups for phenolic content, antioxidant parameters, and IC₅₀ values. Results were considered statistically significant at p < 0.05. All experiments were conducted in triplicate and data are expressed as mean ± standard deviation (SD).

Results and discussions

Optimization of extraction conditions

In our study, TAS and TOS values ​​of A. campestris and A. bisporus were obtained. The determined values ​​are shown in Table 1.

Table 1.

Antioxidant activity results (TAS and TOS) of Agaricus Campestris and A. bisporus extracts obtained under various extraction conditions used for optimization modeling.

Extraction temperature °C	Extraction time (h)	Extract concentration (mg/mL)	A. campestris		A. bisporus
			TAS (mmol /L)	TOS (µmol /L)	TAS (mmol /L)	TOS (µmol /L)
40	3	0.25	2.707 ± 0.082	7.761 ± 0.085	1.549 ± 0.112	4.289 ± 0.204
		0.5	2.789 ± 0.061	7.549 ± 0.085	1.666 ± 0.059	4.091 ± 0.130
		1	2.854 ± 0.048	7.488 ± 0.055	1.722 ± 0.062	3.977 ± 0.120
		2	2.904 ± 0.046	7.333 ± 0.103	1.882 ± 0.045	3.789 ± 0.091
	5	0.25	2.976 ± 0.056	7.019 ± 0.082	1.817 ± 0.101	3.568 ± 0.116
		0.5	3.130 ± 0.063	6.957 ± 0.080	1.922 ± 0.050	3.025 ± 0.690
		1	3.151 ± 0.059	6.900 ± 0.066	1.996 ± 0.046	3.317 ± 0.037
		2	3.238 ± 0.108	6.762 ± 0.073	2.142 ± 0.072	3.081 ± 0.059
	7	0.25	2.237 ± 0.109	7.402 ± 0.051	1.039 ± 0.068	3.894 ± 0.176
		0.5	2.292 ± 0.084	7.578 ± 0.120	1.206 ± 0.040	4.123 ± 0.170
		1	2.333 ± 0.053	7.725 ± 0.038	1.299 ± 0.046	4.262 ± 0.234
		2	2.488 ± 0.093	7.823 ± 0.068	1.460 ± 0.059	4.623 ± 0.054
	9	0.25	1.755 ± 0.102	8.323 ± 0.118	0.942 ± 0.063	4.448 ± 0.110
		0.5	1.871 ± 0.048	8.498 ± 0.072	1.012 ± 0.062	4.559 ± 0.081
		1	1.955 ± 0.039	8.601 ± 0.063	1.184 ± 0.037	4.673 ± 0.105
		2	2.050 ± 0.060	8.756 ± 0.095	1.269 ± 0.088	4.824 ± 0.042
50	3	0.25	4.087 ± 0.177	6.437 ± 0.089	2.371 ± 0.051	2.129 ± 0.167
		0.5	4.236 ± 0.062	6.274 ± 0.059	2.498 ± 0.047	1.935 ± 0.067
		1	4.326 ± 0.037	6.201 ± 0.056	2.606 ± 0.045	1.896 ± 0.051
		2	4.431 ± 0.073	6.082 ± 0.094	2.753 ± 0.045	1.787 ± 0.059
	5	0.25	4.208 ± 0.060	6.037 ± 0.059	2.493 ± 0.116	1.911 ± 0.152
		0.5	4.340 ± 0.105	5.949 ± 0.092	2.644 ± 0.045	1.829 ± 0.116
		1	4.492 ± 0.065	5.840 ± 0.079	2.720 ± 0.034	1.751 ± 0.129
		2	4.690 ± 0.061	5.552 ± 0.082	2.795 ± 0.035	1.573 ± 0.048
	7	0.25	3.469 ± 0.098	6.000 ± 0.093	2.010 ± 0.063	1.809 ± 0.055
		0.5	3.579 ± 0.046	6.114 ± 0.060	2.082 ± 0.080	1.924 ± 0.071
		1	3.649 ± 0.021	6.262 ± 0.099	2.181 ± 0.049	2.042 ± 0.055
		2	3.748 ± 0.038	6.426 ± 0.092	2.283 ± 0.054	2.338 ± 0.064
	9	0.25	2.696 ± 0.033	6.572 ± 0.063	1.823 ± 0.068	2.514 ± 0.026
		0.5	2.819 ± 0.031	6.641 ± 0.081	1.938 ± 0.038	2.669 ± 0.062
		1	2.947 ± 0.045	6.789 ± 0.030	2.010 ± 0.071	2.746 ± 0.049
		2	3.062 ± 0.085	6.929 ± 0.130	2.177 ± 0.159	2.870 ± 0.074
60	3	0.25	3.662 ± 0.092	6.740 ± 0.078	1.991 ± 0.037	2.763 ± 0.059
		0.5	3.802 ± 0.028	6.673 ± 0.082	2.160 ± 0.049	2.664 ± 0.057
		1	3.919 ± 0.028	6.571 ± 0.041	2.289 ± 0.034	2.632 ± 0.092
		2	3.990 ± 0.079	6.472 ± 0.069	2.491 ± 0.065	2.487 ± 0.041
	5	0.25	3.817 ± 0.063	6.398 ± 0.069	2.059 ± 0.079	2.141 ± 0.127
		0.5	3.923 ± 0.055	6.287 ± 0.073	2.209 ± 0.023	2.012 ± 0.037
		1	4.040 ± 0.047	6.160 ± 0.068	2.357 ± 0.060	1.912 ± 0.066
		2	4.237 ± 0.073	6.001 ± 0.071	2.538 ± 0.067	1.774 ± 0.072
	7	0.25	3.039 ± 0.121	6.431 ± 0.122	1.450 ± 0.081	2.412 ± 0.068
		0.5	3.158 ± 0.058	6.550 ± 0.076	1.671 ± 0.039	2.574 ± 0.066
		1	3.250 ± 0.056	6.642 ± 0.109	1.721 ± 0.028	2.772 ± 0.085
		2	3.482 ± 0.062	6.826 ± 0.095	1.948 ± 0.103	2.929 ± 0.100
	9	0.25	1.930 ± 0.074	6.984 ± 0.132	1.172 ± 0.113	2.723 ± 0.089
		0.5	2.063 ± 0.062	7.080 ± 0.070	1.263 ± 0.042	2.903 ± 0.083
		1	2.200 ± 0.042	7.268 ± 0.075	1.441 ± 0.042	2.984 ± 0.029
		2	2.395 ± 0.042	7.468 ± 0.128	1.578 ± 0.039	3.101 ± 0.066
70	3	0.25	2.042 ± 0.091	7.216 ± 0.090	1.315 ± 0.090	3.649 ± 0.066
		0.5	2.241 ± 0.086	7.062 ± 0.078	1.467 ± 0.073	3.568 ± 0.055
		1	2.478 ± 0.084	6.954 ± 0.124	1.652 ± 0.087	3.340 ± 0.142
		2	2.615 ± 0.030	6.638 ± 0.114	1.794 ± 0.072	3.157 ± 0.046
	5	0.25	2.335 ± 0.088	6.526 ± 0.070	1.469 ± 0.059	2.688 ± 0.067
		0.5	2.510 ± 0.036	6.428 ± 0.092	1.634 ± 0.054	2.621 ± 0.028
		1	2.673 ± 0.037	6.312 ± 0.033	1.794 ± 0.054	2.525 ± 0.047
		2	2.799 ± 0.043	6.123 ± 0.080	1.919 ± 0.039	2.405 ± 0.079
	7	0.25	1.743 ± 0.057	6.861 ± 0.111	0.917 ± 0.100	3.062 ± 0.121
		0.5	1.851 ± 0.037	7.143 ± 0.087	1.029 ± 0.059	3.216 ± 0.096
		1	1.960 ± 0.030	7.212 ± 0.041	1.229 ± 0.039	3.344 ± 0.108
		2	2.079 ± 0.114	7.464 ± 0.074	1.403 ± 0.062	3.837 ± 0.121
	9	0.25	0.963 ± 0.123	7.509 ± 0.050	0.948 ± 0.081	3.592 ± 0.072
		0.5	1.107 ± 0.067	7.642 ± 0.044	1.190 ± 0.078	3.743 ± 0.053
		1	1.223 ± 0.043	7.814 ± 0.061	1.269 ± 0.010	3.907 ± 0.076
		2	1.337 ± 0.034	7.933 ± 0.097	1.321 ± 0.030	4.044 ± 0.067

The data reveal a non-linear relationship between extraction parameters and outcomes, with 50 °C as an optimal threshold for A. campestris and A. bisporus, beyond which TAS declines and TOS rises (e.g., 1.337 ± 0.034 mmol/L TAS and 7.933 ± 0.097 µmol/L TOS at 70 °C, 9 h, 2 mg/mL for A. campestris). A 5-hour duration maximizes TAS and minimizes TOS (e.g., 1.573 ± 0.048 µmol/L for A. bisporus at 50 °C, 2 mg/mL), while higher concentrations enhance TAS and reduce TOS. A. campestris consistently shows higher TAS, and A. bisporus lower TOS, highlighting species-specific profiles. These findings support tailored optimization and further modeling.

Among the developed models, the one yielding TAS as the output was selected as the best for single-objective optimization, while the model producing both TAS and TOS outputs was chosen for multi-objective optimization. For A. campestris, the optimal model architecture with TAS as the output was identified as 3-4-1, whereas the best model with both TAS and TOS outputs was determined to be 3-6-2. Similarly, for Agaricus bisporus, the architectures were found to be 3-8-1 for the TAS output model and 3-12-2 for the model with both TAS and TOS outputs. These model architectures were selected based on their performance metrics. For instance, the 3-4-1 architecture for A. campestris (TAS output) yielded the lowest MSE (0.002), highest correlation coefficient R (0.999), and lowest MAPE (1.294). Similarly, the 3-8-1 model for A. bisporus (TAS output) demonstrated the best fit with MSE of 0.002, R of 0.997, and MAPE of 2.058. Multi-output models (e.g., 3-6-2 and 3-12-2) were also chosen according to a balance of MSE and R values across both TAS and TOS outputs. The performance indicators of the selected best models are presented in Table 2.

Table 2.

The performance indicators of the selected best models.

Mushroom	Single-objective optimization results		Multi-objective optimization results
	A.campestris	A.bisporus	A.campestris		A.bisporus
Output	TAS	TAS	TAS	TOS	TAS	TOS
MSE	0.002	0.002	0.002	0.005	0.002	0.004
MAPE	1.294	2.058	1.811	0.678	1.811	1.651
R	0.999	0.997	0.997	0.995	0.997	0.997

The table presents the performance metrics (MSE, MAPE, R) of single- and multi-objective optimization models for A. campestris and A. bisporus. R values (0.995–0.999) indicate excellent fit, while MSE (0.002–0.005) reflects low error. In single-objective optimization, MAPE for TAS is 1.294 for A. campestris and 2.058 for A. bisporus; in multi-objective, MAPE ranges from 0.678 to 1.811 for TAS and TOS. The models demonstrate reliable and effective performance for both species.

For A. campestris mushroom, the single-objective optimization curve and the Pareto set generated during the multi-objective optimization process are illustrated in Fig. 2.

Fig. 2.

The single-objective optimization curve and the pareto set of A. campestris.

In the case of A. campestris, the optimum extraction conditions from single-objective optimization were determined as 50.002 °C temperature, 4.615 h duration, and 1.952 mg/mL extract concentration. For multi-objective optimization, the conditions were established as 51.591 °C temperature, 6.141 h duration, and 1.769 mg/mL extract concentration.

The single-objective optimization graph for A. bisporus mushroom and the Pareto set obtained from the multi-objective optimization process are presented in Fig. 3.

Fig. 3.

The single-objective optimization graph and the pareto set of A. bisporus.

In the single-objective optimization study conducted for A. bisporus, the optimal extraction conditions were determined to be 49.459 °C temperature, 4.983 h duration, and 1.960 mg/mL extract concentration. The multi-objective optimization study resulted in optimum values of 54.359 °C temperature, 7.355 h duration, and 1.683 mg/mL extract concentration.

Antiproliferative activity

While the incidence of cancer cases is increasing today, many different methods such as surgery, chemotherapy, radiotherapy, hormone therapy and immunotherapies are used in the fight against this disease¹⁵. The effectiveness of cancer treatment is directly related to strengthening the immune system. Supporting the immune system helps the body to be more effective against cancer cells during the treatment process¹⁶. For this reason, the use of various supplements and natural products that support the immune system is widely preferred to contribute to the treatment process¹⁷. In our study, the effect of extracts of A. campestris and A. bisporus produced under optimum conditions against the A549 cancer cell line was investigated. The findings are shown in Fig. 4.

Fig. 4.

Antiproliferative activity of A. campestris and A. bisporus optimized extracts. (Control: The group not treated with chemicals, only kept in the medium; DMSO: The group in which the medium and DMSO were applied; the group in which the extract was applied at 25, 50, 100 and 200 µg/mL concentrations; AC-RSM extract: Agaricus campestris RSM optimized extract; AC-ANN-GA extract: Agaricus campestris ANN-GA optimized extract; AB-RSM extract: Agaricus bisporus RSM optimized extract; AB-ANN-GA extract: Agaricus bisporus ANN-GA optimized extract)

In our study, the antiproliferative effects of extracts of A. campestris and A. bisporus produced under optimum conditions on A549 lung cancer cell line was investigated. Figure 4 shows the effects of extracts applied at different concentrations (25, 50, 100 and 200 µg/mL) on cell proliferation. When compared to the control group (cells kept only in the medium) and DMSO group (cells applied with DMSO together with the medium), it was observed that all extracts decreased cell proliferation in a dose-dependent manner. There were significant differences between A. campestris RSM optimized extract (AC-RSM) and A. campestris ANN-GA optimized extract (AC-ANN-GA) and A. bisporus RSM optimized extract (AB-RSM) and A. bisporus ANN-GA optimized extract (AB-ANN-GA). In general, it is noteworthy that extracts optimized with ANN-GA exhibited higher antiproliferative activity compared to RSM optimized extracts. It was determined that all extracts significantly decreased cell viability at the highest concentration level of 200 µg/mL. These findings reveal that optimized extracts of A. campestris and A. bisporus can exhibit antiproliferative effects on A549 lung cancer cells. In particular, extracts optimized with ANN-GA method had a stronger antiproliferative effect compared to those optimized with RSM method, indicating that artificial neural networks (ANN) and genetic algorithms (GA) can be effective tools in the optimization of biological compounds. These methods can contribute to making the bioactive components of mushroom extracts more effective.

Previous studies have reported that Agaricus species contain anticarcinogenic components such as phenolic compounds, polysaccharides and triterpenoids¹⁸. These components have been shown to be effective in cancer cells through mechanisms such as stopping the cell cycle, inducing apoptosis and increasing oxidative stress¹⁹. Our study also shows that A. campestris and A. bisporus suppress cell proliferation in the A549 cell line in a dose-dependent manner, and these findings are consistent with studies in the literature. Previous studies have indicated that Agaricus bisporus is particularly rich in polysaccharide content and can exhibit anticancer effects by modulating the immune system²⁰. However, in our study, it was observed that ANN-GA optimization increased the antiproliferative activity for both A. bisporus and A. campestris. This suggests that optimization processes can enhance the biological activity of extracts and that new generation biotechnological approaches can play an important role in the development of anticancer agents. Although molecular mechanisms were not directly investigated in this study, the observed biological effects may be linked to intracellular pathways. Phenolic compounds such as gallic acid and caffeic acid are known to influence cell viability by modulating apoptotic signaling, possibly through the regulation of pro- and anti-apoptotic proteins or activation of caspases¹⁸. In terms of neuroprotective activity, inhibition of acetylcholinesterase may help preserve synaptic acetylcholine levels, contributing to improved neuronal communication. These mechanistic aspects should be addressed in future studies to better understand the pharmacological potential of mushroom extracts.

In our study, the antiproliferative effects of Agaricus campestris and Agaricus bisporus extracts on A549 lung cancer cell line were investigated and it was determined that ANN-GA optimization increased this effect. In the literature, it has been shown that Agaricus lanipes extracts also inhibited proliferation by inducing apoptosis in A549 cells, especially increased Bax and Caspase-3 gene expressions, and decreased Bcl-2 levels²¹. In a different study, it was shown that the water extract of Agaricus blazei Murill caused growth inhibition and apoptotic cell death in A549 cells, increased p21 expression by arresting the cell cycle in the G2/M phase, and decreased COX-2 levels²². This supports that Agaricus species may have an effect on cancer cells through apoptotic mechanisms. Agaricus species can suppress cancer cells by disrupting the cell cycle and increasing oxidative stress because they contain polyphenols and polysaccharides with high antioxidant capacity²³. The current findings indicate that the results obtained with the optimization methods in our study may increase the therapeutic potential of Agaricus extracts. However, in vivo studies and mechanistic analyses are required. In particular, determining the effects on intracellular signaling pathways and investigating their synergistic effects with chemotherapy agents are important to evaluate the usability of these mushroom species as supportive agents in cancer treatment. In conclusion, it was determined that the anticancer effect of A. campestris and A. bisporus extracts, especially with ANN-GA optimization, can be increased and the bioactive components of these species have pharmaceutical potential. Future studies should be aimed at evaluating the effects of these extracts on intracellular signaling pathways and their synergistic potential with chemotherapy agents.

Anticholinesterase activity

Oxidative stress plays an important role in the development of many chronic diseases, and Alzheimer’s disease is the leading one among them. This disease, which is more common in individuals aged 65 and over, is becoming increasingly common with the aging population^24,25. It is estimated that the number of Alzheimer’s patients worldwide may exceed 80 million in the coming years. The use of antioxidant supplements to prevent the harmful effects caused by oxidative stress may help slow down the progression of the disease²⁶. In this context, the antioxidant activities of A. campestris and A. bisporus extracts produced under optimized conditions were evaluated. The findings are shown in Table 3.

Table 3.

Anti-AChE and anti-BChE values of A. campestris and A. bisporus.

	Optimized Extracts	AChE µg/mL	BChE µg/mL
A. campestris	Single-objective optimization	52.957 ± 0.710b	72.667 ± 0.973b
	Multi-objective optimization	67.400 ± 0.801c	89.747 ± 0.901c
A. bisporus	Single-objective optimization	69.983 ± 0.806d	114.620 ± 1.337d
	Multi -objective optimization	86.837 ± 0.634e	130.060 ± 0.614e
Galantamine		6.367 ± 0.158a	15.453 ± 0.126a

^aMeans having the different superscript letter(s) in the same column are significantly different (p < 0.05) according to Duncan Test (ANOVA).

In our study, the inhibitory effects of extracts obtained from A. campestris and A. bisporus mushroom species under different production conditions on acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) were evaluated via IC₅₀ values. Since the IC₅₀ value expresses the concentration of a compound required to inhibit the activity of the enzyme by 50%, lower IC₅₀ values ​​indicate stronger inhibitory activity²⁷. Our results revealed that in both mushroom species, extracts produced by the single-objective method had lower IC₅₀ values ​​compared to those produced by the multi-objective method and therefore showed stronger anticholinesterase activity. Especially, single-objective extract of A. campestris exhibited the strongest inhibitory activity with IC₅₀ values ​​of 52.957 ± 0.710 µg/mL for AChE and 72.667 ± 0.973 µg/mL for BChE. In general, lower IC₅₀ values ​​of extracts of A. campestris compared to A. bisporus suggest that this mushroom species has higher inhibitory potential. However, inhibitory activities of mushroom extracts are quite weak compared to galantamine used as a pharmaceutical agent. Galantamine has very low IC₅₀ values ​​of 6.367 ± 0.158 µg/mL for AChE and 15.453 ± 0.126 µg/mL for BChE and is a much more effective compound in inhibiting enzymes compared to mushroom extracts. These results show that mushroom extracts alone may not be sufficient in the treatment of neurodegenerative diseases, but they have potential as natural inhibitor sources. The findings obtained reveal that production conditions may have a significant effect on the bioactivity of mushroom extracts. The lower IC₅₀ values ​​of extracts produced with the single-objective method suggest that this method may strengthen the inhibitory activity by increasing the amount of phenolic compounds, polysaccharides and other bioactive components. In this context, further studies are needed to examine the effects of different production techniques on bioactive compound profiles in more detail. It has been previously reported in the literature that A. bisporus exhibits inhibitory properties of acetylcholinesterase and butyrylcholinesterase enzymes^28,29. However, in our study, it was observed that A. bisporus exhibited lower inhibitory activity compared to A. campestris. This suggests that different cultivation conditions and extraction methods may directly affect the amounts of bioactive compounds and inhibitory capacities. Studies reported in the literature reveal that A. bisporus is a mushroom rich in phenolic compounds and polysaccharides, but the amount of these components may vary depending on the production methods. In particular, the growth environment, harvest time and extraction methods used may play an important role in different results of inhibitory activity. In addition, anticholinesterase activities of mushroom extracts can be evaluated as potential natural inhibitors in the treatment of Alzheimer’s and other neurodegenerative diseases. However, due to high IC₅₀ values, purification, component isolation and bioavailability studies should be carried out to make these compounds more effective pharmacologically. Although the inhibitory capacity of A. bisporus is consistent with previous studies, the results obtained show that production and extraction conditions are a determining factor. It is thought that mushroom extracts can be used as supportive agents in combination therapies that show synergistic effects rather than being strong inhibitors alone. Future studies should evaluate the effects of different cultivation and extraction methods on inhibitory capacity more comprehensively.

Antioxidant activity

Mushrooms are natural resources known for their strong antioxidant properties and can have significant effects on health. Thanks to the bioactive compounds they contain, they can have protective effects on cellular health by reducing oxidative damage caused by free radicals in the body^30,31. In addition, the potential of mushrooms to strengthen the immune system and provide resistance to various diseases is also being investigated. These properties make mushrooms valuable as natural health supporters³². In our study In our study, the antioxidant activities of the extracts of A. campestris and A. bisporus produced under optimum conditions were determined. The findings are shown in Table 4.

Table 4.

Antioxidant values of A. campestris and A. bisporus.

Mushroom	Optimization	TAS (mmol/L)	DPPH (mg Trolox Equi/g)	FRAP (mg Trolox Equi/g)	TOS (µmol/L)	OSI (TOS/(TAS*10))
A. campestris	Single-objective	5.017 ± 0.011d	74.030 ± 0.724d	93.048 ± 0.964d	8.273 ± 0.017d	0.165 ± 0.001d
	Multi-objective	4.167 ± 0.019c	62.246 ± 0.913c	78.867 ± 0.903c	5.758 ± 0.019c	0.138 ± 0.001c
A. bisporus	Single-objective	2.821 ± 0.018b	47.246 ± 0.754b	60.947 ± 0.736b	3.163 ± 0.030b	0.112 ± 0.002b
	Multi-objective	2.075 ± 0.026a	41.762 ± 1.338a	51.887 ± 0.730a	1.479 ± 0.007a	0.071 ± 0.001a

^aMeans having the different superscript letter(s) in the same column are significantly different (p < 0.05) according to Duncan Test (ANOVA).

The antioxidant activities of the extracts of two different mushroom species, A. campestris and A. bisporus, produced under optimum conditions were evaluated and the results are presented in Table 4. According to the data obtained, A. campestris extracts showed higher antioxidant activity compared to A. bisporus extracts. While the A. campestris extract produced by the single target optimization method had a TAS value of 5.017 mmol/L, this value decreased to 4.167 mmol/L in the multi-target optimization. Similarly, these values ​​for A. bisporus were determined as 2.821 mmol/L and 2.075 mmol/L, respectively. DPPH and FRAP analyses used to determine the antioxidant capacity also showed a similar trend. A. campestris extracts had DPPH values ​​of 74.030 mg Trolox Equivalent/g (single-target) and 62.246 mg Trolox Equivalent/g (multi-target), while these values ​​for A. bisporus were measured as 47.246 mg Trolox Equivalent/g and 41.762 mg Trolox Equivalent/g, respectively. FRAP analysis also shows that the antioxidant capacity of A. campestris is higher than A. bisporus. In the literature, the TAS value of A. bisporus was reported as 1.256 mmol/L, the TOS value as 11.473 µmol/L, and the OSI value as 0.915³³. Compared to these studies, it was determined that both the single-objective and multi-objective optimized extracts of A. bisporus used in our study had higher TAS values. In addition, TOS value and OSI value were determined to be lower in the samples used in our study. In addition, TAS value of wild mushrooms Cantharellus cibarius was reported as 5.511 mmol/L, TOS value as 7.289 µmol/L, OSI value as 0.132³⁴. TAS value of Lactarius deliciosus was reported as 7.468 mmol/L, TOS value as 13.161 µmol/L, OSI value as 0.176³⁵. TAS value of Hericium erinaceus was reported as 5.426 mmol/L, TOS value as 6.621 µmol/L, OSI value as 0.122¹⁴. Compared to these studies, it was determined that both single-objective and multi-objective optimized extracts of A. campestris and A. bisporus used in our study had lower TAS values. In addition, it was determined that the TOS value of the single-objective optimized extract of A.campestris was higher than Cantharellus cibarius and Hericium erinaceus, and lower than Lactarius deliciosus. It was observed that the multi-objective optimized extract of A. campestris had lower TOS values ​​than Cantharellus cibarius, Lactarius deliciosus and Hericium erinaceus. It was observed that the OSI values ​​of both single-objective and multi-objective optimized extracts of A. campestris were higher than Cantharellus cibarius and Hericium erinaceus, and lower than Lactarius deliciosus. It was observed that TOS and OSI values ​​of both single-objective and multi-objective optimized extracts of A. bisporus were lower than Cantharellus cibarius, Lactarius deliciosus and Hericium erinaceus. The higher antioxidant activity of A. campestris compared to A. bisporus may be due to differences in phytochemical components between species. The concentration of polyphenols, flavonoids and other antioxidant components may vary depending on the type of mushroom, growing conditions and extraction method³⁶. This situation is especially more pronounced in extracts obtained with the single-target optimization method, and it is seen that multi-target optimization causes a slight decrease in antioxidant capacity. The fact that TAS, TOS and OSI values ​​are lower compared to wild mushrooms reported in the literature suggests that the growth environments, substrate composition and genetic characteristics of the mushrooms are effective on their antioxidant potential. The exposure of wild mushrooms to harsher environmental conditions may have triggered their defense mechanisms to produce more antioxidant compounds. In addition, the fact that the TOS and OSI values ​​of Agaricus species are generally lower than wild species indicates that their total oxidative load is relatively lower, and therefore their oxidative stress levels are low. In addition, when the effect of optimization methods on antioxidant activity is considered, it is seen that single-target optimization provides higher antioxidant capacity. This shows that while focusing on a single parameter ensures maximum antioxidant compounds, multi-target optimization is more balanced but may cause some loss in antioxidant capacity. In conclusion, these data reveal that optimization strategies aimed at increasing the antioxidant properties of mushroom extracts should be carefully selected. Also, while multi-objective optimization provides a balanced outcome, it involves inherent trade-offs. In our study, extracts optimized through multi-objective optimization generally exhibited lower TAS, DPPH, and FRAP values compared to single-objective optimization, due to the simultaneous effort to minimize TOS. This reflects a shift from maximizing a single target to balancing multiple outcomes. Such trade-offs are expected in multi-objective optimization, where improving one objective (e.g., reducing oxidant levels) can limit the maximization of another (e.g., TAS or phenolic content). This highlights the importance of aligning optimization goals with application priorities, such as targeting higher antioxidant power versus lower oxidative stress.

Phenolic contents

Mushrooms synthesize various bioactive compounds to protect against environmental stress. These compounds strengthen the defense mechanisms against harmful microorganisms and external factors by ensuring the survival of the fungi. In particular, compounds with antioxidant and anticancer properties help fungi play important roles in their ecosystems in nature^1,37. These biological activities also make mushrooms valuable natural resources in the field of health. In our study, the phenolic contents of the extracts of A. campestris and A. bisporus produced under optimum conditions were determined in the LC-MS/MS device. The findings obtained are shown in Table 5.

Table 5.

Phenolic contents of A. campestris and A. bisporus.

Phenolic compounds	A. campestris		A. bisporus
	Single-objective	Multi-objective	Single-objective	Multi-objective
Acetohydroxamic acid	744.380 ± 1.390d	583.800 ± 1.100c	122.200 ± 1.280b	84.470 ± 0.860a
Gallic acid	6761.850 ± 6.360d	6025.260 ± 11.630c	3198.85 ± 3.70b	2966.650 ± 2.210a
Protocatechuic acid	4169.400 ± 1.670c	4244.570 ± 2.430d	1024.31 ± 2.310b	939.540 ± 1.950a
4-hydroxybenzoic acid	None	none	1825.450 ± 2.480b	1249.600 ± 2.990a
Caffeic acid	7242.030 ± 1.930d	6177.580 ± 4.570c	2356.790 ± 2.660b	2325.410 ± 2.320a
Quercetin	1694.990 ± 2.730d	1188.650 ± 1.180c	648.050 ± 1.720b	603.590 ± 1.690a
Catechinhyrate	545.970 ± 1.860c	693.270 ± 1.060d	432.930 ± 0.930b	402.970 ± 1.410a
Myricetin	None	none	558.660 ± 2.340b	469.720 ± 1.630a
Vanillic acid	130.300 ± 0.840b	124.920 ± 0.790a	None	none

^aMeans having the different superscript letter(s) in the same line are significantly different (p < 0.05) according to Duncan Test (ANOVA).

The results of the phenolic compound analysis show that there are significant differences between A. campestris and A. bisporus species. In general, it was determined that phenolic compounds were found at higher concentrations in A. campestris extracts. In particular, gallic acid, protocatechuic acid and caffeic acid were detected in both mushroom species but were found at significantly higher levels in A. campestris. This may be one of the main factors explaining the higher antioxidant capacity of A. campestris than A. bisporus. 4-hydroxybenzoic acid and myricetin were determined only in A. bisporus extracts, and the presence of these compounds reflects the phytochemical differences between the species. The findings obtained in this study are also consistent with previous studies. For example, it was reported that phenolic compounds such as gallic acid and catechin were found in the analyzes conducted in A. bisporus species³⁸. In this study, it was observed that the phenolic compound levels of A. campestris and A. bisporus species were lower compared to some wild mushroom species, which suggests that genetic differences within the region and species may affect phenolic compound synthesis. Vanillic acid, in particular, was detected only in A. campestris extracts. Some previous studies reported low levels of vanillic acid in different mushroom species³⁹. However, the fact that vanillic acid could not be detected in A. bisporus can be explained by the difference in the phenolic metabolism of this species. These results suggest that the biosynthetic pathways of mushrooms may vary among species and that the distribution of phenolic compounds may differ depending on the genetic structure of the mushroom, the conditions of the region where it is collected, and the extraction methods^40,41. When evaluated in terms of optimization methods, it was observed that the levels of phenolic compounds were generally higher in the extracts produced with the single-target optimization method. This shows that single-target optimization, which aims to obtain a certain phenolic compound at the maximum level, may be a more effective method compared to multi-target optimization. However, the catechinhydrate compound was detected at higher levels in A. campestris extracts obtained with the multi-target optimization method. This suggests that the synthesis mechanism of certain compounds may be sensitive to the extraction conditions used⁴². Similarly, it has been emphasized that the phenolic content of mushroom extracts may vary with extraction parameters and that optimum conditions should be determined for the preservation of phenolic compounds⁴². In conclusion, this study reveals that the phenolic content of A. campestris is richer than A. bisporus and that this has an effect on antioxidant activity. In addition, it has been determined that the phenolic compound content is sensitive to optimization strategies, single-target optimization generally produces extracts with higher phenolic content, but some specific compounds can be found at higher levels under different optimization conditions. It has been revealed that in the evaluation of mushrooms for functional food and pharmaceutical applications, phenolic compound profiles should be examined in detail and optimization strategies should be determined accordingly. The phenolic compound profile of A. campestris showed significantly higher concentrations of gallic acid, protocatechuic acid, and caffeic acid compared to A. bisporus. The phenolic compound profile of A. campestris revealed significantly higher levels of gallic acid, protocatechuic acid, and caffeic acid compared to A. bisporus. This difference can be attributed both to the species’ intrinsic metabolic capacity and the efficiency of the ANN-GA optimization approach used in the study. The optimized extraction conditions likely enhanced the recovery of these phenolic compounds, allowing A. campestris to exhibit its full phytochemical potential. In addition, A. campestris may possess a higher biosynthetic ability to produce phenolic compounds in response to environmental factors such as oxidative stress or nutrient availability. Genetic and metabolic differences between the two species may also influence the activity of enzymes involved in phenolic biosynthesis, further contributing to the observed variation in phenolic content.

Conclusion

Optimum extraction conditions for obtaining bioactive compounds from A. campestris and A. bisporus were identified, and their biological activities were comprehensively evaluated. The optimization process performed using ANN and GA provided maximum bioactive compounds. The results show that ANN-GA optimization method is effective in producing extracts with higher biological activity compared to traditional methods. Analyses revealed that A. campestris has higher total antioxidant capacity (TAS) and lower total oxidant level (TOS) compared to A. bisporus. These strong antioxidant properties may be attributed to the higher phenolic content, particularly gallic acid, which is also associated with enhanced antiproliferative and anticholinesterase activities observed in A. campestris. In addition, A. campestris has higher anticholinesterase activity and it has been shown that it may be a potential natural inhibitor source especially for neurodegenerative diseases such as Alzheimer’s. Antiproliferative activity tests showed that the extracts obtained by ANN-GA optimization suppressed cell proliferation in A549 lung cancer cell line. These findings indicate the usability of Agaricus species in functional foods, food supplements, and pharmaceutical products. In future studies, the effects of mushroom extracts on intracellular molecular mechanisms should be investigated in more detail. In particular, studies on signaling pathways, gene expression levels, and apoptosis mechanisms will provide a better understanding of the effects of bioactive compounds in cancer and neurodegenerative diseases. In addition, the safety, bioavailability, and efficacy of mushroom extracts should be tested with in vivo animal models and clinical studies. In addition, the stability and bioavailability of bioactive components of mushroom extracts can be increased by using nanotechnology-based carrier systems (e.g. liposomes, nanoparticles). In addition, it is recommended to conduct a comprehensive optimization study using comparative analyses of different Agaricus species and different extraction methods. In contrast to previous optimization studies, our research integrates ANN-GA modeling with broader biological validation, including antiproliferative and anticholinesterase evaluations, which underscores its novelty and potential pharmaceutical significance. The superiority of the ANN-GA approach over RSM in this study can be attributed to its ability to model complex, nonlinear relationships between multiple extraction parameters and biological outcomes. Unlike RSM, which relies on predefined second-order polynomial models, ANN can flexibly learn from data without assuming any specific mathematical form. The integration with GA further enhances this capability by efficiently exploring the solution space and avoiding local minima, thus improving the convergence toward global optimal conditions. These methodological advantages enable ANN-GA to better capture the multifactorial dynamics underlying the extraction process and result in more accurate predictions and superior biological activities. In conclusion, this study emphasizes the importance of artificial intelligence-supported optimization techniques to increase the bioactive components of Agaricus species and shows that these species have high potential in the biotechnology, food, and pharmaceutical industries. One limitation of this study is the lack of testing on normal (non-cancerous) cell lines. Future studies should include comparisons with normal cell lines (e.g., HEK293 or MRC-5) to evaluate the selectivity and safety profile of the extracts. Future studies will be aimed at better understanding the clinical and pharmaceutical use potential of these fungal species.

Author contributions

A.G. and M.S. wrote the main manuscript text and prepared all of figures . All authors reviewed the manuscript.

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.