Optimization of Hymenopellis radicata extracts using RSM and ANN-GA and evaluation of biological activities

Abstract

Currently, the utilization of bioactive chemicals derived from natural products is on the rise across various sectors, including pharmaceuticals and the food industry. Our investigation identified the extraction procedures that yield the most biological activity of the medicinal fungus Hymenopellis radicata. The antioxidant, anticholinesterase, and antiproliferative properties of the extracts generated under these conditions were assessed. In this context, temperature, duration, and solvent ratio served as optimization parameters, which were obtained utilizing Response Surface Methodology (RSM) and Artificial Neural Network-Genetic Algorithm (ANN-GA). The antioxidant capacities of the optimized extracts were assessed using TAS, TOS, DPPH, and FRAP methodologies. Anticholinesterase activity were evaluated through the inhibition of AChE and BChE. The antiproliferative activity was assessed using the MTT assay on A549 (lung), MCF-7 (breast), and DU-145 (prostate) cancer cell lines. Analyses of phenolic content were conducted by LC-MS/MS. Analyses indicated that extracts optimized using RSM demonstrated superior biological activity relative to those optimized with ANN-GA. They were additionally discovered to possess decreased TOS values. Moreover, studies of phenolic content indicated elevated compound levels in extracts optimized by RSM. Thus, H. radicata was identified as a significant natural resource from both pharmaceutical and biotechnological viewpoints. Moreover, optimization tactics were identified as a crucial factor directly influencing biological activity.

Introduction

Medicinally valuable mushrooms are being studied with increasing interest not only in traditional medicine but also in modern science. These organisms are notable for their rich nutritional content. They constitute an important natural source of high-quality protein, digestible dietary fiber, essential amino acids, vitamins (especially B complex, vitamins C and D), and minerals (e.g., potassium, selenium, magnesium, and zinc)^1,2. In addition to being suitable for healthy nutrition with their low fat and calorie content, they also possess various pharmacological effects thanks to their biologically active components. The polysaccharides and β-glucans they contain have immunomodulatory effects, phenolic compounds and flavonoids have strong antioxidant properties, and triterpenes and sterols exhibit antimicrobial and anticancer activities^3,4. Recent studies have revealed that medicinal mushrooms support cardiovascular health, protect nerve cells against oxidative damage, and may play a role in the prevention of metabolic diseases. Considering both their nutritional value and biological activities, these mushrooms stand out as a unique natural resource for the development of functional food and pharmaceutical products in modern health approaches^5,6. Optimizing the extract conditions of medicinal mushrooms is a critical process for the most efficient extraction of polysaccharides, phenolics, triterpenes, and other bioactive compounds found in these organisms. Extraction efficiency varies depending on factors such as solvent type, temperature, time, and solid-liquid ratio. In addition to the widely used Soxhlet extraction, modern techniques such as ultrasonic, microwave, and supercritical fluid assisted extraction are also employed. This makes it possible to develop valuable mushroom extracts for functional food and pharmaceutical products that are standardized on both scientific and industrial scales^7–10.

H. radicata (formerly Xerula radicata) is a macrofungus distributed in different ecological zones worldwide, attracting attention with its taxonomic diversity and nutritional properties. It belongs to the Physalacriaceae family. New varieties have been identified, particularly in regions such as India and Japan, and their morphological and microscopic characteristics have been documented in detail^11,12. In addition, its widespread occurrence in different regions indicates that it is an important component of macrofungal diversity^13,14. It has also been reported that both the mycelium and fruit body contain high amino acids, making it suitable for use as a functional food. Its richness in protein and amino acids is one of the most important characteristics that enhance the species’ nutritional value¹⁵. No published studies on the phenolic composition of H. radicata have been found in the literature. However, numerous investigations on basidiomycete species have shown that mushrooms generally contain a broad spectrum of phenolic compounds, including gallic acid, protocatechuic acid, syringic acid, vanillic acid, caffeic acid, and p-hydroxybenzoic acid^16–18. These phenolics contribute significantly to the antioxidant, antimicrobial, and cytoprotective properties commonly observed in edible and medicinal mushrooms. In many species, phenolic acids function as potent radical scavengers, metal chelators, and modulators of oxidative pathways, underscoring their importance in the development of functional extracts^16–19. Therefore, although the phenolic profile of H. radicata remains unknown, it is reasonable to assume that it may follow biochemical patterns similar to other basidiomycetes and warrants investigation using optimized extraction approaches. Recent studies have increasingly utilized chemometric and AI-assisted methods to optimize mushroom extraction processes and enhance their biological activity. AI-assisted optimization significantly improved the biological activities of Phylloporia ribis²⁰, while both RSM and ANN models successfully predicted antioxidant, anticholinesterase, and antiproliferative responses in Paralepista flaccida²¹. Similar approaches applied to Agaricus species²² and Lactarius deliciosus²³ also demonstrated the effectiveness of single- and multi-objective optimization in enhancing biological activity profiles. Despite these advancements, no studies have investigated the optimization of extraction parameters or compared the predictive performance of RSM and ANN-GA for H. radicata. Thus, the present study fills a notable gap by integrating dual optimization models with comprehensive biological activity evaluation, representing a novel application of RSM and ANN-GA to an underexplored basidiomycete. Therefore, the purpose of the present study is to optimize the extraction parameters of H. radicata using RSM and ANN-GA and to evaluate the optimized extracts in terms of their antioxidant, antimicrobial, antiproliferative, anticholinesterase activities, and phenolic contents. By establishing predictive models and identifying conditions that maximize biological activity, this study provides the first comprehensive assessment of the bioactive potential of H. radicata and contributes valuable insights into the therapeutic potential of unexplored basidiomycetes.

Materials and methods

The mushroom specimens utilized in this research were collected during field studies conducted in the Kastamonu region of Türkiye (41.3885° N, 33.7816° E). The taxonomic identification was verified by comparison with standard mycological keys, and voucher specimens were deposited in the Ankara University Fungarium under the accession number MS-360. These details ensure traceability and reproducibility of the collected materials.

Extraction procedure

Our study assessed temperature, extraction duration, and solvent-to-water ratio as experimental variables to identify the optimal extraction conditions. All extraction procedures were conducted with an adjustable ultrasonic bath apparatus. Twenty-seven distinct experimental settings were established based on these experimental parameters. To achieve this objective, various extraction conditions were established: temperatures of 30, 45, and 60 °C; durations of 30, 45, and 60 min; and solvent ratios of 0, 50, and 100%. The TAS values of the extracts acquired under these 27 circumstances were assessed utilizing RSM. Additionally, optimization utilizing artificial intelligence was conducted through the integration of ANN and GA.

Response surface methodology (RSM)

RSM was applied to determine the optimal extraction conditions using extraction temperature, time, and ethanol/water ratio as independent variables, while Total Antioxidant Status (TAS) served as the response. The optimization was performed with Design Expert 13 software, and the data were analyzed using a second-order polynomial regression model. The mathematical framework employed throughout the modeling phase was predicated on the generic phrase provided below:

In the model, Y_k represents the TAS value, while X_i corresponds to the coded independent variables (temperature, time, and ethanol/water ratio). The constant term is expressed as β_ko. Model validity was evaluated through the coefficient of determination (R²), ANOVA, and p-values. Derivative analysis was used to determine the optimal extraction conditions, and three-dimensional response surface plots illustrated the interactions among variables.

Artificial neural network-genetic algorithm (ANN-GA)

The ANN-GA hybrid model was developed to predict and optimise the Total Antioxidant Status (TAS, mmol/L) based on three input variables: extraction temperature (°C), extraction time (min), and ethanol percentage (%). The dataset, consisting of 27 experimental observations obtained from the RSM design, was divided into training (80%), validation (10%), and testing (10%) subsets to enhance model precision and generalization. The Artificial Neural Network (ANN) was trained using the Levenberg-Marquardt (LM) algorithm due to its high convergence efficiency and stability. To prevent overfitting, a 5-fold cross-validation procedure was employed. Various hidden-layer neuron numbers (1–20) were tested to balance model complexity and computational cost, with the best predictive accuracy obtained for a four-neuron hidden layer, establishing the final 3-4-1 architecture (three input neurons, one hidden layer with four neurons, and one output neuron). The ANN parameters were set as follows: learning rate = 0.5, momentum = 0.5, maximum iterations = 500, validation stop = 50, and error tolerance = 1 × 10⁻⁵. The Genetic Algorithm (GA) was implemented in MATLAB R2023b (MathWorks Inc., Natick, MA, USA) using custom scripts. It was configured with a population size of 50, a crossover rate of 0.7, a mutation rate of 0.02, and 100 generations. The GA’s objective function was defined as the maximisation of the ANN-predicted TAS values to identify the optimal combination of extraction parameters yielding the highest antioxidant response. The final 3-4-1 ANN model achieved strong predictive performance, with Mean Square Error (MSE) = 0.0028, Mean Absolute Percentage Error (MAPE) = 3.72%, and correlation coefficient (R) = 0.982, confirming excellent agreement between experimental and predicted results.

1

2

In the equations, e_i represents the experimental values, p_i denotes the predicted values from the ANN model, and n is the total number of observations. GA was applied in the optimization phase, with its performance evaluated under different population sizes. New individuals were generated using a roulette wheel selection and single-point crossover to maintain genetic diversity. Optimization efficiency was examined through convergence graphs, and each run was repeated 30 times to increase the likelihood of achieving the global optimum.

Extraction for bioactivity

This study identified the optimal extraction parameters to enhance the biological activity of extracts derived from mushroom samples. Optimization experiments utilizing RSM identified optimal conditions: a temperature of 42.786 °C, an extraction duration of 40.930 min, and a 71.789% ethanol/water ratio. The ANN-GA hybrid model forecasted ideal parameters of 38.435 °C, 50.286 min, and a 52.317% ethanol/water ratio. Extraction processes were conducted according to the parameters most closely aligned with those derived from both methodologies. An ultrasonic bath apparatus (WIGGENS UA22MFDN, Germany) operating at a frequency of 40 kHz, with full power at 100% and a capacity of 400 W, was utilized in this procedure.

Antioxidant activity assays

The total antioxidant status of optimized mushroom extracts was assessed using the TAS Assay Kit (Rel Assay Diagnostics, Mega Tıp, Gaziantep, Turkey; Product Code: RL0017). The TAS method relies on the reduction of the 2,2′-azinobis(3-ethylbenzothiazoline-6-sulfonate) (ABTS) radical cation by antioxidant chemicals found in the extracts. The color decrease from the reaction was quantified spectrophotometrically and represented as mmol Trolox equivalent/L²⁴. The total oxidant status was evaluated utilizing the TOS Assay Kit (Product Code: RL0024) provided by the same company. The TOS test operates on the idea that oxidizing chemicals change iron ions to their ferric state, which subsequently form complexes with xylenol orange. The reaction’s color intensity was quantified spectrophotometrically, with results reported as µmol hydrogen peroxide equivalent/L²⁵. The Oxidative Stress Index (OSI), indicative of the oxidative equilibrium of the samples, was determined by assessing the combined values of TAS and TOS. The OSI was calculated by dividing the TOS by the TAS, with the results represented as a percentage (%). This measure was deemed a significant indicator that quantitatively demonstrates the equilibrium between the antioxidant capacity and oxidant load of the extracts²⁶.

The free radical scavenging capacity of the optimized mushroom extracts was determined by the DPPH (2,2-diphenyl-1-picrylhydrazyl) assay. Stock solutions (1 mg/mL) were prepared in dimethylsulfoxide (DMSO, Sigma-Aldrich, St. Louis, MO, USA). For each test, 1 mL of extract solution was mixed with 160 µL of 0.267 mM DPPH and 4 mL of 0.004% methanol (Merck, Darmstadt, Germany). The mixture was incubated in the dark at room temperature for 30 min, and absorbance was measured at 517 nm using a UV-VIS spectrophotometer. Results were expressed as mg Trolox equivalent per gram of extract (mg TE/g)²⁷.

The ferric ion reducing power of the optimized mushroom extracts was measured using the FRAP (Ferric Reducing Antioxidant Power) assay. For each test, 100 µL of extract stock solution was mixed with 2 mL of freshly prepared FRAP reagent composed of 300 mM acetate buffer (pH 3.6), 20 mM FeCl₃·6 H₂O in 40 mM HCl, and 10 mM TPTZ solution (Sigma-Aldrich, St. Louis, MO, USA) in a 10:1:1 ratio. The mixtures were incubated at 37 °C for 4 min, and absorbance was recorded at 593 nm using a UV-Vis spectrophotometer. Results were expressed as mg Trolox equivalents per g of extract (mg TE/g)²⁷.

Anticholinesterase activity assays

The anticholinesterase effects of optimized mushroom extracts were assessed by modifying the Ellman technique²⁸. Galantamine hydrobromide (Sigma-Aldrich, St. Louis, MO, USA) served as the reference inhibitor in the tests. Extracts were produced at quantities between 200 and 3.125 µg/mL. In microplate applications, 130 µL of 0.1 M phosphate buffer (pH 8.0, Merck, Darmstadt, Germany) was initially dispensed into each well. Subsequently, 20 µL of acetylcholinesterase (AChE, Type VI-S, EC 3.1.1.7) or butyrylcholinesterase (BChE, EC 3.1.1.8) enzyme solutions (Sigma-Aldrich, St. Louis, MO, USA) were incorporated, accompanied by 10 µL of extract solution. The resulting mixes were incubated for 10 min at 25 °C in darkness to facilitate reaction. Subsequent to the incubation period, 20 µL of DTNB [5,5′-dithiobis-(2-nitrobenzoic acid), Sigma-Aldrich, St. Louis, MO, USA] was introduced to each well to commence the reaction. Twenty microliters of substrate solution (acetylcholine iodide or butyrylcholine iodide, Sigma-Aldrich, St. Louis, MO, USA) were included into the solution. The extent of enzyme inhibition was assessed using absorbance measurements with a microplate reader set to 412 nm. Results are presented as IC₅₀ values in µg/mL.

Antiproliferative activity assays

The antiproliferative activity of the optimized mushroom extracts was tested against three human cancer cell lines: A549 (lung adenocarcinoma, ATCC CCL-185), MCF-7 (breast adenocarcinoma, ATCC HTB-22), and DU-145 (prostate cancer, ATCC HTB-81), all obtained from the American Type Culture Collection (ATCC, USA). Extract concentrations of 25, 50, 100, and 200 µg/mL were applied to cells cultured under standard conditions. Upon reaching 70–80% confluence, cells were trypsinized (3 mL Trypsin-EDTA; Sigma-Aldrich, USA) and seeded in plates for 24 h pre-incubation. Extracts were then added for another 24 h, followed by treatment with 1 mg/mL MTT solution [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; Sigma-Aldrich, USA] at 37 °C until purple formazan crystals formed. Crystals were dissolved in DMSO (negative control), and absorbance was measured at 570 nm using an Epoch spectrophotometer (BioTek Instruments, USA). The results confirmed the antiproliferative efficacy of the optimized extracts²⁷.

Phenolic analysis

The phenolic chemical profiles of the optimized mushroom extracts were analyzed with a high-sensitivity LC-MS/MS technology. The presence and quantities of 24 distinct standard phenolic compounds were assessed. An analytical C18 Intersil ODS-4 column (3.0 mm×100 mm, 2 μm; GL Sciences, Tokyo, Japan) was employed for chromatographic separation, with the column temperature consistently maintained at 40 °C during the analysis. The mobile phase system was formulated using two distinct solutions: Phase A consisted of ultrapure water (Milli-Q, Millipore, Bedford, MA, USA) with 0.1% formic acid (Merck, Darmstadt, Germany); Phase B comprised LC-MS grade methanol (Merck, Darmstadt, Germany) enriched with 0.1% formic acid. The flow rate for the investigation was established at 0.3 mL/min, with a sample volume of 2 µL injected into the system for each instance.

Statistical analysis

The statistical analyses of the experimental data from this investigation were conducted using the SPSS 21.0 software (IBM Corp., Armonk, NY, USA). An Independent Samples t-test was employed to analyze the mean differences between two independent groups. One-way analysis of variance (One-Way ANOVA) was employed to compare three or more groups. An alpha level of 0.05 was employed to interpret the ANOVA results. The Duncan Multiple Range Test was employed to identify the groups responsible for major variations among them.

Results and discussions

Extraction optimization

This study meticulously analyzed three critical extraction parameters that affect TAS readings. The independent variables in the study included extraction temperature (30, 45, and 60 °C), application duration (30, 45, and 60 min), and solvent composition (0%, 50%, and 100% ethanol/water ratio). Table 1 presents the TAS values derived from extractions conducted with various combinations of these parameters at distinct levels.

Table 1.

TAS values ​​of the extracts obtained in the study.

Experiment number	Ethanol/water ratio (%)	Extraction time (min)	Extraction temperature (°C)	TAS (mmol/L)
1	0	30	30	5.542 ± 0.023j
2	0	45	30	5.846 ± 0.031m
3	0	60	30	4.731 ± 0.027c
4	50	30	30	5.863 ± 0.019m
5	50	45	30	6.065 ± 0.025o
6	50	60	30	4.889 ± 0.017e
7	100	30	30	5.653 ± 0.029k
8	100	45	30	5.954 ± 0.023n
9	100	60	30	4.831 ± 0.016d
10	0	30	45	5.744 ± 0.012l
11	0	45	45	5.947 ± 0.027n
12	0	60	45	4.956 ± 0.023f
13	50	30	45	5.890 ± 0.021m
14	50	45	45	6.158 ± 0.032p
15	50	60	45	5.069 ± 0.020g
16	100	30	45	5.842 ± 0.019m
17	100	45	45	6.048 ± 0.027o
18	100	60	45	4.848 ± 0.017de
19	0	30	60	5.258 ± 0.021h
20	0	45	60	5.439 ± 0.034i
21	0	60	60	4.549 ± 0.019a
22	50	30	60	5.463 ± 0.019i
23	50	45	60	5.537 ± 0.023j
24	50	60	60	4.656 ± 0.021b
25	100	30	60	5.277 ± 0.018h
26	100	45	60	5.470 ± 0.018i
27	100	60	60	4.575 ± 0.049a

^aMeans having the different superscript letter(s) in the same column are significantly different (p < 0.05) according to Duncan’s multiple range test.

The TAS data obtained in this study reveal that the delicate balance between extraction parameters directly affects biological activity. The findings indicate that extraction performed at 45 °C, for 45 min, and with a solvent containing 50% ethanol yielded the highest TAS value of 6.158 ± 0.032 mmol/L. On the other hand, the lowest TAS result (4.848 ± 0.017 mmol/L) obtained with 100% ethanol and extension of the extraction time to 60 min shows that the stability of the antioxidant components decreases with increasing extraction time. In particular, the generally low TAS values measured at 60 °C (range: 4.549 ± 0.019–5.537 ± 0.023 mmol/L) confirm that high temperatures have a negative effect on the antioxidant structure. Furthermore, systematic decreases observed under conditions where the extraction time was extended to 60 min occurred similarly at all temperature levels, indicating that this time exceeded the limiting value. On the other hand, the results obtained at 30 °C exhibited a more consistent distribution, with a decrease recorded between the highest value (6.065 ± 0.025 mmol/L; 45 min, 50% ethanol) and the lowest value (4.731 ± 0.027 mmol/L; 60 min, 0% ethanol), although this decrease was not as pronounced as at higher temperatures. The study also employed two different optimization approaches based on the experimental findings.

Within the RSM framework, four distinct regression models were compared: linear, two-factor interaction (2FI), quadratic, and cubic. The R² score of 0.984 signifies that the quadratic model offers the most robust fit to the data set. This result signifies that 98.4% of the variance in the dependent variable is elucidated by the independent variables, indicating that the model possesses substantial explanatory power. The elevated coefficient of determination indicates that the model effectively captures the relationships between the independent variables and the response variable, accurately reflecting the overarching trends in the observational data. Consequently, it was determined that the quadratic model yields reliable and highly predictive results in optimization analyses for extraction operations. The quadratic polynomial regression model developed to forecast TAS values for the H. radicata species is delineated below:

In the defined quadratic regression model, X₁ denotes the extraction temperature, X₂ signifies the extraction time, and X₃ indicates the ethanol/water ratio. The three-dimensional response surface plots illustrating the individual and interactive effects of these parameters on TAS values derived from H. radicata extracts are displayed in Fig. 1.

Fig. 1.

Response surface plots.

In the second part of the study, the ANN algorithm was utilized to predictively model experimental findings. As a result of comprehensive performance analyses conducted on different network architectures, the structure with the highest predictive accuracy was selected, and an advanced optimization approach was developed by integrating this structure with GA. In this context, the ANN model, designed with a 3-4-1 architecture and containing four neurons in a single hidden layer, was identified as the most suitable structure considering statistical success criteria. The model’s predictive capacity was verified with values ​​of MSE value 0.008, MAPE value 0.377%, and R value 0.998. These high accuracy levels demonstrate that the selected network architecture can be used as a reliable and effective optimization tool. The GA-based optimization process continued the search for a solution using the most reliable prediction outputs obtained from the ANN model as a reference. At this stage, population size, a critical variable that directly affects the algorithm’s performance, was tested and compared at different values. Based on these evaluations, the optimal number of individuals was determined to be 10. The convergence graph presented in Fig. 2 clearly shows that the fitness value increased rapidly during the initial iterations and stabilised after approximately nine iterations, indicating that the Genetic Algorithm successfully converged to an optimal and balanced solution.

Fig. 2.

Convergence graph.

Antioxidant activity

Mushrooms are natural sources with significant antioxidant potential. Thanks to the compounds they contain, they have a high potential to reduce oxidative stress. These compounds protect the body from cellular damage by reducing the effects of free radicals. In this context, determining the antioxidant potential of mushrooms is crucial^29,30. In our study, we tested the antioxidant potential of optimized extracts of H. radicata, and the results are presented in Table 2.

Table 2.

Antioxidant values of Hymenopellis radicata optimized extract.

Parameters	RSM extract values	ANN-GA extract values
FRAP (mg Trolox Equi/g)	160.047 ± 2.386a	146.787 ± 3.405b
TAS (mmol/L)	6.308 ± 0.057a	6.078 ± 0.065b
DPPH (mg Trolox Equi/g)	122.997 ± 2.603a	94.837 ± 2.552b
TOS (µmol/L)	4.320 ± 0.035b	5.662 ± 0.056a
OSI (TOS/(TAS*10))	0.068 ± 0.001b	0.093 ± 0.001a

In this study, the antioxidant potentials of optimized extracts of H. radicata were determined using the RSM and ANN-GA methods. The results showed that RSM extracts had higher FRAP and DPPH activities. Consequently, the reducing power and free radical scavenging potential of RSM extracts were determined to be higher. Furthermore, TAS values were found to be relatively higher in the RSM extract. This demonstrated that the total antioxidant capacity was more advantageous with the RSM approach. Conversely, TOS and OSI values were found to be higher in extracts optimized with the ANN-GA method. These findings suggest that although the ANN-GA method better captures complex and nonlinear structures in modeling extraction conditions, it creates a higher oxidative load in the extracts in this particular case.

The obtained results clearly demonstrate that the choice of optimization approach can directly affect biological activity results. In particular, better optimization of reducing capacity-based parameters such as FRAP and DPPH using RSM indicates that statistical modeling is more efficient in the extraction of polyphenols and similar antioxidant compounds. In the literature, the antioxidant capacity of H. radicata has been determined using DPPH and FRAP tests and has generally been reported to be moderate³¹. Compared to this study, the findings obtained in our study indicate that the high antioxidant values ​​of extracts obtained using the RSM method can be increased by determining appropriate extraction conditions. Furthermore, no previous findings in the literature have been found regarding the TAS, TOS, and OSI values ​​of Hymenopellis radicata. Research involving various macrofungus species has indicated a TAS value of 4.285 mmol/L, a TOS value of 8.616 µmol/L, and an OSI value of 0.201 for Macrolepiota mastoidea³². The TAS of Cantharellus cibarius was recorded at 5.511 mmol/L, the TOS at 7.289 µmol/L, and the OSI at 0.132³³. The TAS of Hericium erinaceus was documented as 5.426 mmol/L, the TOS as 6.621 µmol/L, and the OSI as 0.122³⁴. The TAS of Paralepista flaccida was recorded at 4.054 mmol/L, the TOS at 10.352 µmol/L, and the OSI at 0.255²¹. In comparison to these investigations, it was noted that both the RSM and ANN-GA extracts of H. radicata utilized in our research exhibited elevated TAS values relative to M. mastoidea, C. cibarius, H. erinaceus, and P. flaccida. The TAS number signifies the total antioxidant chemicals present in natural products. The TOS number signifies the total oxidant chemicals generated in natural goods. The OSI value indicates the percentage reduction of oxidant molecules by antioxidant compounds⁹. The RSM and ANN-GA extracts of H. radicata exhibited decreased TOS and OSI values compared to M. mastoidea, C. cibarius, H. erinaceus, and P. flaccida. This comparative assessment reveals how the methods used to optimize the extraction conditions of H. radicata shape its biological activity. The TAS, TOS, and OSI values, reported for the first time in the literature, indicate that the species has a significantly stronger antioxidant capacity compared to other medically important macrofungal species. The higher TAS values ​​in extracts optimized with both RSM and ANN-GA compared to other mushroom species reported in the literature demonstrate that H. radicata is rich in phenolic and similar antioxidant compounds, and optimization techniques play a critical role in unlocking this potential. The lower TOS and OSI values ​​compared to other species demonstrate not only high antioxidant capacity but also successful suppression of oxidant compounds in the extracts. In comparison with other medicinal mushrooms, the TAS, DPPH, and FRAP values obtained for H. radicata were notably higher than those reported for H. erinaceus and C. cibarius. Moreover, the DPPH scavenging activity observed in this study approaches that of the synthetic antioxidant Trolox under similar experimental conditions. These comparative results indicate that the antioxidant potential of H. radicata is not only statistically significant but also practically comparable to that of well-established natural and synthetic antioxidants, underscoring its biopharmaceutical relevance.

The striking point here is the differential impact of the optimization approach used on the results. The RSM method reduced the oxidative load with lower OSI values while maintaining high TAS values. However, the ANN-GA method, despite producing high TAS, produced relatively higher OSI values along with higher TOS. This demonstrates that, despite ANN-GA’s success in modeling nonlinear relationships, it does not always provide the most appropriate output in terms of biological function. Therefore, determining the optimization strategy for H. radicata should be aligned not only with efficiency or modeling accuracy, but also with the biological functionality of the extract. These findings support the need to prioritize biological activity targets when selecting optimization methods for natural product extraction, and that statistical approaches may yield more advantageous results in certain species. The higher oxidative load (TOS and OSI) observed in ANN-GA extracts, despite their elevated TAS values, may be attributed to differences in extraction kinetics and compound stability under the algorithm’s optimized conditions. The ANN-GA approach, which focuses on global data fitting rather than local thermodynamic behavior, may have favored parameters that enhance total extraction yield but also promote the co-extraction of unstable or easily oxidizable compounds. In contrast, the statistically guided RSM method likely provided milder extraction conditions that preserved antioxidant integrity while minimizing oxidation-prone molecule release. Furthermore, the ANN-GA optimization may have exhibited less selectivity toward specific phenolic subclasses, resulting in a broader but less stable compound profile.

Anticholinesterase activity

Extracts from diverse mushroom species exhibit inhibitory effects on cholinesterase enzymes. The inhibition of acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) is particularly notable as a possible target for the treatment of neurodegenerative illnesses³⁵. Consequently, natural fungal-derived inhibitors are regarded as significant prospects in pharmaceutical research for neurological illnesses, including Alzheimer’s and Parkinson’s³⁶. The anticholinesterase activity of optimized extracts of H. radicata was assessed in our investigation. The acquired IC₅₀ values are displayed in Table 3.

Table 3.

Anticholinesterase activity of Hymenopellis radicata optimized extract.

Sample	AChE µg/mL	BChE µg/mL
Galantamine	6.60 ± 0.25c	16.19 ± 0.35c
ANN-GA extract	101.93 ± 2.18a	137.15 ± 2.20a
RSM extract	80.98 ± 1.43b	121.93 ± 2.13b

In our study, we investigated the anticholinesterase activities of optimized extracts of H. radicata. It was determined that extracts obtained using both optimization methods provided significant inhibition of both AChE and BChE enzymes. However, compared to the standard inhibitor galantamine, the inhibitory potency of mushroom extracts was found to be quite low. This can be explained by the fact that natural products generally contain a combination of multiple phenolic compounds and secondary metabolites rather than a single compound, resulting in weaker and broader-spectrum enzyme inhibition³⁷. Anticholinesterase activity of different mushroom species has been reported^38–40. In our study, we determined that extracts optimized with ANN-GA exhibited lower inhibitory activity, exhibiting higher IC50 values ​​on both AChE (101.93 µg/mL) and BChE (137.15 µg/mL). In contrast, extracts optimized with the RSM method achieved lower IC50 values ​​(80.98 µg/mL for AChE and 121.93 µg/mL for BChE), thus demonstrating a greater advantage in terms of enzyme inhibition. This finding suggests that the statistically based RSM approach contributes to more efficient extraction of anticholinesterase compounds. Furthermore, it is understood that while algorithmic models such as ANN-GA can successfully model nonlinear relationships between complex parameters, they do not always provide the optimal conditions for biological activity. However, RSM, with its lower IC50 values ​​and stronger inhibition profile, provides more balanced conditions for the solubility and efficient extraction of phenolics and flavonoids that determine anticholinesterase activity in mushroom extracts. These results demonstrate that the selection of an optimization method for the pharmacological functionalization of natural products should focus not only on modeling success but also on the level of targeted biological activity.

Antiproliferative activity

Some mushroom species with edible and medicinal properties are attracting attention with their anticancer potential. The polysaccharides, triterpenes, phenolic compounds, and other bioactive metabolites contained in these mushrooms exert potent effects on cancer cells by suppressing cell proliferation, inducing apoptotic mechanisms, and limiting metastatic processes^41,42. Mushrooms are natural items distinguished by their anticancer effects⁴³. This study assessed the antiproliferative effects of optimized extracts of H. radicata on A549, MCF-7, and DU-145 cancer cell lines. The results are illustrated in Fig. 3.

Fig. 3.

Antiproliferative activity of Hymenopellis radicata optimized extract (Control group: Cells were maintained only in standard culture medium, with no added chemicals. DMSO group: To assess the solvent effect, only DMSO was applied to the cells, with no extract added. Experimental groups: Cells were treated with extract solutions prepared at concentrations of 25, 50, 100, and 200 µg/mL, respectively).

In our study, the antiproliferative activities of extracts optimized with RSM and ANN-GA from H. radicata were tested against A549, MCF-7, and DU-145 cell lines. At the end of the study, it was determined that there was a decrease in cell viability with increasing concentrations of both RSM and ANN-GA optimized extracts. While cell viability was high in the control and DMSO groups, it was observed that extracts applied at concentrations of 100 and 200 µg/mL suppressed cell proliferation. In this context, it was observed that H. radicata extracts have anticancer potential in different cancer cell lines. It was determined that extracts obtained with RSM exhibited stronger antiproliferative activities than extracts obtained with ANN-GA. RSM extracts caused a more significant decrease in viability, especially in the MCF-7 cell line. Similarly, it was observed that they exhibited higher cytotoxic effects with lower optical density (OD) values ​​in A549 and DU-145 cells. Although an antiproliferative effect was observed in extracts optimized with ANN-GA, this effect was determined to be lower than that obtained with the RSM method. These results are thought to be due to the more favorable conditions under which optimal solubility and bioavailability of phenolic compounds and other biologically active molecules are achieved with RSM. In the literature, H. radicata has been reported to exhibit higher selectivity against the MDA-MB-231 breast cancer cell line. Furthermore, it has been reported that it does not produce a significant antiproliferative effect in the MCF-7 cell line³¹. Compared to this study, the extracts optimized with RSM in our study also showed significant antiproliferative activity in the MCF-7 cell line. This is thought to be due to the optimization approach used, which contributed to the different results obtained from previous findings in the literature by increasing the bioavailability of the active compounds in the extract. Consequently, it was observed that the optimization method can determine not only the extraction efficiency but also the anticancer activity profile. Although the IC₅₀ values obtained for H. radicata extracts indicate moderate antiproliferative potency compared with synthetic chemotherapeutics, such activity levels are considered promising for natural crude extracts. Many clinically relevant anticancer agents, such as paclitaxel and doxorubicin, initially exhibited similar cytotoxic ranges in early in vitro screenings before purification and formulation optimization. Therefore, the observed activity suggests that H. radicata possesses bioactive compounds with potential pharmacological value that merit further isolation and mechanistic evaluation. In the present study, cell viability assays were conducted using standard controls, including serum-free (SF) medium and DMSO, both of which exhibited high viability rates. The reduced viability observed in the extract-treated cells compared to these controls clearly reflects the cytotoxic and antiproliferative potential of the extracts. Since the SF and DMSO controls already served as baseline viability references, an additional normal cell line was not deemed necessary for comparative purposes in this design. Nevertheless, future studies will extend these findings by incorporating normal fibroblast or epithelial cells to further assess the selectivity index and safety profile of H. radicata extracts.

Phenolic contents

Mushrooms provide as significant natural sources of phenolic chemicals. The bioactive chemicals they generate fluctuate based on the employed extraction parameters⁴⁴. The phenolic contents of optimal extracts of H. radicata were quantified using an LC-MS/MS apparatus in our investigation. The results are presented in Table 4.

Table 4.

Phenolic contents of Hymenopellis radicata optimized extract.

Phenolic compounds	Retention time	LOD	LOQ	ANN-GA extract (mg/kg)	RSM extract (mg/kg)
2-hydroxycinnamic acid	2.907	7.79	25.98	264.22 ± 2.66b	694.41 ± 2.11a
Protocatechuic acid	0.548	7.17	23.90	4014.30 ± 1.96b	4363.78 ± 2.43a
Gallic acid	0.429	22.88	76.25	12930.46 ± 3.00b	15297.67 ± 3.89a
Quercetin	4.768	68.4	228.1	1209.74 ± 2.17b	1238.85 ± 2.30a
Catechinhydrate	0.549	7.64	25.47	1525.49 ± 2.27b	1732.62 ± 2.84a

In our study, the phenolic contents of optimized extracts of H. radicata were determined. The levels of phenolic compounds varied depending on the optimization method used. The amounts of gallic acid, protocatechuic acid, catechin hydrate, and 2-hydroxycinnamic acid in the extract optimized with the RSM method were found to be significantly higher than those obtained with the ANN-GA method. In particular, the three-fold higher level of 2-hydroxycinnamic acid in the RSM extract demonstrates that extraction conditions directly affect the solubility and bioavailability of phenolic compounds. Although quercetin levels were similar in both methods, RSM was found to provide a richer overall phenolic profile. These findings are of great importance given the strong relationship between phenolic compounds and biological activities. Compounds such as gallic acid and protocatechuic acid contribute to high antioxidant capacity^45,46, while flavonoids such as catechin and quercetin stand out with their potential for both antiproliferative and antimicrobial effects^47,48. In this context, the stronger antioxidant (FRAP, DPPH, TAS) and antiproliferative activity performance of extracts obtained with the RSM method is directly related to the richness of their phenolic contents. The ANN-GA method was more limited in the total yield of phenolic compounds, which was reflected in the relatively lower levels of biological activity obtained. Furthermore, despite the ability of ANN-GA to model nonlinear relationships, RSM appears to be more advantageous in the efficient extraction of phenolic compounds that enhance biological activity. Therefore, in studies aimed at revealing the biological potential of natural products, not only theoretical modeling success but also the acquisition of phenolic metabolites at functional levels should be the primary criterion. In this study, RSM’s higher optimization of phenolic compounds clearly reveals that it offers a more suitable approach to increase the antioxidant and antiproliferative potential of H. radicata. To further clarify the relationship between phenolic compounds and biological activities, Pearson correlation analysis was performed between major phenolics (gallic acid, protocatechuic acid, catechin, quercetin, and 2-hydroxycinnamic acid) and antioxidant parameters (TAS, FRAP, DPPH) as well as antiproliferative activity. Strong positive correlations (p < 0.01) were observed, indicating that higher phenolic contents, particularly gallic and protocatechuic acids, were closely associated with increased antioxidant capacity and cytotoxic potential. These findings statistically support the mechanistic link proposed between phenolic composition and bioactivity.

Conclusion

This study identified the optimal extraction procedures that yield the maximum biological activity of H. radicata. The biological activity of the extracts generated under optimal circumstances were assessed. It should be emphasised that the present study does not claim methodological superiority of the ANN–GA model over RSM; rather, the extracts optimised through ANN-GA exhibited higher biological activity, reflecting compositional differences under the respective optimal conditions. The investigations revealed that the biological activity of the extracts varied according to the optimization settings. The extracts derived from the RSM approach demonstrated superior activity relative to those obtained via the ANN-GA method. RSM extracts demonstrated elevated FRAP, TAS, and DPPH activity. Furthermore, it was established that the activities of both AChE and BChE were elevated in the extracts procured by the RSM approach. The extracts generated under optimal circumstances from the mushroom demonstrate considerable potential as natural inhibitors for the treatment of neurodegenerative disorders. RSM extracts had significant antiproliferative effects against A549, MCF-7, and DU-145 cell lines. In this setting, the mushroom was discovered to potentially serve as a significant anticancer agent. LC-MS/MS analysis indicated that phenolic compounds, including gallic acid, protocatechuic acid, catechin, and quercetin, were present in elevated amounts and exhibited enhanced biological activities in the RSM method, suggesting a direct influence of phenolic compounds on activity. Thus, it was concluded that extracts of H. radicata procured under ideal conditions demonstrated considerable biological activity. Moreover, several optimization circumstances were discovered to directly influence biological activities. Quantitative evaluation revealed that the phenolic composition, particularly the elevated levels of gallic and protocatechuic acids, strongly correlated with enhanced antioxidant and antiproliferative activities. These findings confirm that optimized extraction conditions, especially those established via RSM, not only improve the yield of functional phenolic compounds but also directly determine the biological efficacy of H. radicata extracts. Thus, phenolic concentration serves as a reliable biochemical indicator of extract functionality and optimization success.

Author contributions

Conceptualization, A.G., M.S, T.K., A.F.K., E.C.E., I.A.; methodology, A.G., M.S, A.F.K., I.A.; validation, A.G., M.S, T.K., A.F.K., E.C.E., I.A.; investigation, A.G., M.S,; resources, A.G., M.S, T.K.,; data curation, A.G., M.S, T.K., A.F.K., E.C.E., I.A.; writing-original draft preparation, A.G., M.S, T.K., A.F.K., E.C.E., I.A.; writing-review and editing, A.G., M.S, T.K., A.F.K., E.C.E., I.A. All authors have read and agreed to the published version of the manuscript.

Data availability

The datasets generated analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.