Propylene glycol-based green extraction of polyphenols from industrial hemp stems: process optimization and bioactivity assessment

Abstract

See More

Industrial hemp (Cannabis sativa L. subsp. sativa) is a multipurpose crop widely cultivated for its fibers, seeds, and oils. Despite the common use of hemp stems for fiber production in textiles and construction, they are frequently discarded as agricultural waste. This study aimed to enhance the utilization of hemp stems through the optimization of ultrasound-assisted extraction (UAE) employing aqueous propylene glycol solvent system as the extraction solvent, guided by Box–Behnken design (BBD). The optimal extraction parameters—an extraction duration of 30 min, a solvent–solute ratio of 28.25 mL/g, and a PG concentration of 32.72%—resulted in a hemp stem extract (HUPG) enriched with bioactive constituents exhibiting significant antioxidant activity. Following analysis by LC–QTOF–MS/MS, a total of 18 phytochemicals were detected, including isofeuric acid, m-coumaric acid, and chelidonic acid. HUPG exhibited antibacterial activity against Staphylococcus aureus, Streptococcus epidermidis, Escherichia coli, and Pseudomonas aeruginosa, and anti-inflammatory effects through 5-lipoxygenase inhibition and NO radical scavenging (1.41 ± 0.38 mg GAE/g). In LPS-induced RAW 264.7 macrophages, HUPG demonstrated inhibitory effect on NO production. Moreover, it enhanced wound closure in HaCaT cells (51.92 ± 6.05% at 10 mg/mL). These findings highlight the promise of HUPG as a sustainable source of bioactive compounds with potential applications in cosmetic and pharmaceutical formulations.

Graphical Abstract

Background

In recent years, plant-derived agents have gained increasing attention in wound-healing research due to their multifunctional therapeutic properties, including antioxidant, anti-inflammatory, and antimicrobial effects. These properties are particularly valuable in addressing challenges associated with the wound-healing process, which is highly complex and often delayed by infections, prolonged inflammation, and other complications [1]. Although traditional synthetic drugs are widely used, their effectiveness is frequently accompanied by adverse effects and the escalating challenge of antimicrobial resistance (AMR) [2]. Consequently, there is a growing demand for safer, more sustainable alternatives. Plant-based wound-healing agents offer a promising solution by not only mitigating potential side effects but also aligning with environmentally friendly and accessible healthcare practices [1].

Among the many plant sources being investigated for their therapeutic potential, industrial hemp (Cannabis sativa L. subsp. sativa) has emerged as a particularly promising candidate. As a highly versatile crop, hemp is widely used in fiber, seed, and oil production, yet a significant portion of its biomass—particularly the stems—remains underutilized and is often discarded as agricultural waste [3]. Given the increasing interest in sustainable wound-healing agents, hemp stems present an untapped resource with considerable potential. Recent studies have identified them as a rich source of bioactive compounds, including polyphenols (p-coumaric acid and caffeic acid), flavonoids (quercetin and apigenin), and lignans, which exhibit potent antioxidant, antimicrobial, and anti-inflammatory properties [4, 5]. These bioactive constituents are regarded as valuable candidates for innovative developments in pharmaceuticals, nutraceuticals, and cosmetics [6]. However, the lack of efficient strategies to extract and utilize these compounds has limited the valorization of hemp stems. To address this challenge, innovative extraction techniques that align with sustainability principles are needed to maximize bioactive compound recovery from hemp stems while minimizing environmental impact.

In this context, ultrasound-assisted extraction (UAE) has emerged as a promising method for extracting bioactive compounds from plant materials, including hemp stems. UAE utilizes high-frequency sound waves to disrupt plant cell walls, enhancing the release of bioactive compounds into the solvent [7]. In comparison with conventional extraction methods, UAE offers enhanced efficiency through shorter processing times, minimal solvent use, and lower energy input. Its operation under mild thermal conditions also helps preserve the integrity of heat-sensitive phytochemicals. These advantages support UAE as a practical and eco-friendly technique for the extraction and valorization of bioactive compounds from hemp stems [8]. Selecting a suitable solvent system is essential for achieving high extraction yields [9]. In this study, aqueous propylene glycol (PG) was employed as the selected solvent for bioactive compounds extraction owing to its established safety, biocompatibility, and broad regulatory acceptance across pharmaceutical, food, and cosmetic sectors. This binary solvent system synergistically enhances extraction efficiency by leveraging water's polarity and propylene glycol’s solvating capacity, particularly for polar and semi-polar phytochemicals such as phenolic compounds. In addition, its low environmental impact, reusability, and reduced reliance on organic solvents make it consistent with green chemistry principles and economically sustainable extraction practices [10].

Previous studies primarily focused on the extraction of bioactive compounds from hemp leaves, seeds, and flowers [11,12,13,14], leaving the stems relatively unexplored. Despite growing interest in the therapeutic potential of hemp, the extraction of bioactive compounds from its stems remains underexplored, creating a critical gap in the current literature. In response, this study introduces a sustainable approach by applying UAE with aqueous PG solvent system, optimized using Box–Behnken design (BBD), to enhance the recovery of bioactive constituents. The biological efficacy of the hemp stem extracts was systematically assessed through a range of in vitro and cellular assays, covering antioxidant, antibacterial, anti-inflammatory, and wound-healing properties. By coupling green extraction technology with biocompatible solvents, the study presents a promising strategy for developing therapeutic agents from underutilized plant materials. This approach not only supports sustainable resource utilization but also paves the way for eco-friendly innovations in natural product-based wound care.

Materials and methods

Sample extraction and sample preparation for biological assays

Dried stems of C. sativa L. subsp. sativa (hemp) were generously provided by Dr. Natchaya Kuptapan of Samakkee Farm Co. Ltd., located in Sattahip, Chonburi, Thailand. These plant materials were subsequently dried again using a tray dryer set at 45 °C for a period of 3 days [15]. After drying, the stems were milled into a fine powder using a mechanical grinder. The ground material was then passed through a 1 mm mesh sieve to ensure consistent particle size and was stored at ambient conditions until needed.

Extraction of the bioactive constituents was carried out using a modified UAE technique, adapted from Martín-García et al. (2022) [16]. The extraction process was performed using an ultrasonic water bath (Model SAG, Kudos, China) operating at 200 W and holding 10 L of water. Ultrasonic intensity (W/cm²) was determined based on the applied power percentage, total output, and bath volume. For each extraction, 1 g of the hemp powder was mixed with a solvent system according to a predetermined solvent–solute ratio (SSR). Ultrasonic extraction was performed using an intermittent sonication protocol consisting of 5-min treatment intervals alternated with 5-min rest periods, repeated continuously until the predetermined extraction duration was achieved [17]. To control the temperature within the range of 27–32 °C, 7 L of the bath water was routinely replaced with fresh water of the same volume.

Prior to content determination and bioactivity testing, the extracts were appropriately diluted to ensure that their inherent color did not interfere with the colorimetric measurements. The extraction solvent was not removed, as the extract was intended to be directly incorporated into the final formulation, thereby reducing additional processing steps and minimizing time and energy consumption. For all colorimetric assays, distilled water was used as a control, whereas the respective culture media were used as controls for cell-based assays to assess the combined effects of the solvent and bioactive compounds under identical experimental conditions.

Design of the optimization experiment

Preliminary tests were carried out to assess the effects of key variables and potential interacting factors on polyphenol extraction, as detailed in the Supplementary Information. Conditions yielding the highest TPC were selected for subsequent optimization. Extraction duration (ED), SSR, and concentration of solvent (COS) were identified as the most influential variables. The BBD was employed to optimize these parameters using response surface methodology (RSM). Each factor was encoded at three levels: − 1 (low), 0 (medium), and + 1 (high), with actual values listed in Table 1. The design comprised 18 experimental runs, including six center-point replicates (Table 2). Data were modeled using a second-order polynomial equation (Eq. 1):

$${\text{Y = }}\beta_{0} + \sum_{i = 1}^{n} \beta_{i} X_{i} + \sum_{i \ne j = 1}^{n} \beta_{ij} X_{i} X_{j} + \sum_{i = 1}^{n} \beta_{ii} X_{i}^{2}$$

(1)

Table 1 Independent factor coding and experimental levels applied in BBD

Table 2 Comparison of experimental and predicted values from BBD experiments

In the model equation, β₀ represents the intercept, βᵢ denotes the linear coefficient associated with factor Xᵢ, βᵢᵢ is the quadratic coefficient of Xᵢ, and βᵢⱼ corresponds to the interaction coefficient between factors Xᵢ and Xⱼ. These coefficients were defined for n = 3, reflecting the three independent variables assessed in the experimental design.

Assessment of bioactive constituents and antioxidant properties

The total phenolic content (TPC), total flavonoid content (TFC), and total tannin content (TTC) of the samples were determined using colorimetric methods with minor modifications from previously established protocols [18]. TPC was assessed using the Folin–Ciocâlteu method, with absorbance measured at 765 nm and results expressed as mg gallic acid equivalent (GAE)/g sample. TFC was quantified using a sodium nitrite–aluminum chloride colorimetric assay, with absorbance recorded at 510 nm and results expressed as mg quercetin equivalent (QE)/g sample. TTC was determined by reacting the sample with Folin–Ciocâlteu reagent and sodium carbonate, followed by absorbance measurement at 700 nm; results were reported as mg tannic acid equivalent (TAE)/g sample.

Antioxidant activity was evaluated using DPPH radical scavenging and FRAP assays. For the DPPH assay, the sample was reacted with a 0.1 mM DPPH solution and incubated in the dark for 30 min; absorbance was measured at 517 nm, and results were expressed as mg ascorbic acid equivalent (AAE)/g sample. The FRAP assay involved the reaction of the sample with freshly prepared FRAP reagent, followed by incubation for 4 min at room temperature. Absorbance was measured at 593 nm, and antioxidant capacity was expressed as mg AAE/g sample.

Confirmation of optimal parameters

To identify the optimal extraction conditions, numerical optimization was performed based on the goals set for both independent variables and response parameters, as shown in Table 4. The optimization aimed to minimize the independent parameters, simultaneously, the response variables were set to be maximized. Each variable was assigned equal importance (weight = 1), and constraints were applied using the desirability function approach. The optimized conditions were selected based on achieving the highest predicted desirability score, and experimental validation was conducted under these conditions to confirm the model’s predictive accuracy.

Comparison of bioactive compound recovery from hemp using UAE and maceration with selected and common solvents

To comprehensively evaluate extraction efficiency, UAE was systematically compared with conventional maceration using both the optimized solvent system and ethanol. UAE was performed under optimized conditions based on BBD, while maceration was carried out by incubating 1 g of hemp stem powder in each solvent at room temperature for 24 h with continuous shaking at 100 rpm, as modified from a previous study [19].

Chromatographic profiling of phytochemicals of hemp stem extracts

The phytochemical constituents of hemp stem extracts were characterized using liquid chromatography–quadrupole time-of-flight mass spectrometry (LC–QTOF–MS), following the protocol described by Poomanee et al. (2024) [20]. All samples were carefully homogenized, extracted under standardized conditions, and filtered through 0.22 µm membranes before injection. Chromatographic analyses were performed on an Agilent 1290 Infinity II UHPLC system coupled to a 6545B QTOF/MS detector (Agilent Technologies, Santa Clara, CA, USA), equipped with a C18 Poroshell column (2.1 × 150 mm, 2.7 µm particle size), which provides high resolution and efficiency for the separation of complex phytochemical matrices. The mobile phase consisted of acetonitrile (A) and water (B), and a carefully optimized gradient elution program was applied: 5% A/95% B at 1 min, linearly increased to 17% A/83% B from 1 to10 min, held until 13 min, subsequently ramped to 95% A/5% B from 13 to 20 min and maintained until 25 min, then re-equilibrated to initial conditions (5% A/95% B) from 25 to 27 min and held until 33 min. A flow rate of 0.2 mL/min was selected to ensure optimal peak resolution and reproducibility. For mass spectrometric detection, 1 µL of each sample was introduced via multi-wash needle mode to minimize carryover. The LC–QTOF–MS was operated in negative ion mode, which allowed the broad detection of acidic and neutral metabolites. Precursor ions were isolated by a quadrupole mass analyzer and subjected to tandem mass spectrometry (MS/MS) fragmentation according to their mass-to-charge ratio (m/z), providing detailed structural information. The system was operated in time-of-flight (TOF) mode, enabling high-resolution, exact-mass measurements over the range of m/z 50–1100, thus encompassing both low- and high-molecular-weight constituents. Instrumental conditions were optimized for efficient ionization, with the drying gas (N₂) flow rate set to 10 L/min at 300 °C, nebulizer pressure at 35 psi, and nozzle voltage maintained at 1000 V to enhance sensitivity. Spectral acquisition was performed at one spectrum per second, ensuring accurate and reproducible peak detection. Data were processed using Agilent MassHunter Acquisition B.06.01 software, and compound identification was achieved by matching experimental exact masses and MS/MS fragmentation patterns with theoretical values.

Pharmacological assessment of hemp stem extracts

Antibacterial activity assay of hemp stem extracts

The antibacterial potential of hemp stem extracts was investigated against both gram-positive and gram-negative bacterial strains, including Staphylococcus aureus (TISTR 746), Staphylococcus epidermidis (DMST 15505), Escherichia coli (TISTR 527), and Pseudomonas aeruginosa (TISTR 1257). Inoculum preparation involved suspending isolated colonies from overnight-grown Mueller–Hinton agar cultures into sterile 0.85% NaCl solution. The bacterial density was adjusted to correspond with a 0.5 McFarland standard (~ 1.5 × 10⁸ CFU/mL), and accuracy was confirmed using a spectrophotometer at 600 nm [21].

The determination of minimum inhibitory concentration (MIC) followed the broth microdilution technique using 96-well microtiter plates. Hemp extracts were initially prepared at a 50% v/v concentration and serially diluted with Mueller–Hinton broth. Equal volumes of the standardized bacterial suspension and each dilution were added to the wells, which were then incubated at 37 °C for 18 h. To visualize bacterial viability, 0.015% resazurin solution was introduced, and colorimetric changes were monitored after an additional 4-h incubation. The MIC was interpreted as the lowest concentration that inhibited visible bacterial growth, as indicated by the absence of color change [22].

To evaluate bactericidal efficacy, the minimum bactericidal concentration (MBC) was determined by streaking aliquots from wells with no visible growth—including MIC and higher concentrations—onto fresh Mueller–Hinton agar. After incubation at 37 °C for 24 h, the MBC was defined as the lowest concentration that completely inhibited colony formation, indicating bactericidal action. All procedures were performed in duplicate under sterile conditions to ensure data reliability [23].

Cell culture

Antioxidant activity and wound-healing assays were performed using HaCaT human keratinocytes, while RAW 264.7 murine macrophages were used for cellular anti-inflammatory analysis. Each assay was conducted in triplicate to ensure data consistency.

Evaluation of cytotoxic effects and intracellular antioxidant potential of hemp stem extracts

The cytotoxic potential of C. sativa L. subsp. sativa (hemp) stem extracts was assessed in HaCaT and RAW 264.7 cells using the MTT assay [18]. Cells were seeded into 96-well plates at a density of 1.5 × 10⁴ cells per well and cultured until they reached approximately 80% confluence. Cytotoxicity was evaluated by treating the cells with various concentrations of hemp extracts, with ascorbic acid serving as the positive control and serum-free medium as the negative control. After a 24-h incubation period, MTT reagent was added, and cells were incubated for an additional 4 h. The resulting formazan crystals were dissolved in DMSO, and absorbance was measured at 570 nm. Cell viability was calculated as a percentage relative to the untreated control group.

To evaluate cellular antioxidant activity, pre-treated cells were exposed to 300 µM hydrogen peroxide for 1 h to induce oxidative stress. Subsequently, MTT reagent was added and incubated for 4 h. Absorbance was measured at 570 nm, and the protective effect of the extracts was expressed as the percentage of viable cells relative to non-stressed controls. All assays were conducted in triplicate to ensure reproducibility.

In vitro and cellular anti-inflammatory activity assays

The anti-inflammatory activity of the samples was evaluated through both in vitro and cellular assays. Nitric oxide (NO) radical scavenging was determined by the Griess reaction, and 5-lipoxygenase (5-LOX) inhibition was assessed using a modified method from Liana et al. (2024) [24], with nordihydroguaiaretic acid (NDGA) as the positive control and a non-enzymatic background control. Inhibition (%) was calculated as

$${\text{Percent Inhibition }} = \, \left[ {\left( {{\text{X }} - {\text{ Y}}} \right) \, / \, \left( {{\text{X }} - {\text{ Z}}} \right)} \right] \, \times { 1}00,$$

For cellular assessment, RAW 264.7 macrophages were treated with lipopolysaccharide (LPS) to induce NO production, and indomethacin was used as the reference anti-inflammatory agent [25]. The percentage inhibition of NO was calculated using

$${\text{Percent Inhibition }} = \, \left[ {\left( {{\text{X }} - {\text{ Y}}} \right) \, / \, \left( {{\text{X }} - {\text{ Z}}} \right)} \right] \, \times { 1}00,$$

where X refers to the absorbance of LPS-treated cells, Y refers to cells treated with LPS followed by sample treatment, and Z refers to cells treated with LPS followed by control treatment. All assays were performed in triplicate under identical conditions.

Cell scratch assay for wound-healing potential evaluation

The wound-healing potential of the samples was assessed using a scratch assay in HaCaT human keratinocyte cells, following the method described by Kongkadee et al. (2022) [26]. Cells were seeded into 24-well plates at a density of 1.5 × 10⁵ cells/well and cultured to near confluence. A linear scratch was created using a sterile pipette tip, followed by washing with PBS to remove detached cells. Treatment media containing the test extract were added, with complete DMEM serving as the negative control and allantoin at the tested concentration of 200 µg/mL as the positive control. After 18 h of incubation, wound closure was evaluated by comparing the scratched area at 0 and 18 h using ImageJ software. The percentage of wound closure was calculated as

$${\text{Wound closure }}\left( \% \right) \, = \, \left[ {\left( {{\text{Area of wound gap at }}0{\text{h }}{-}{\text{ Area of wound gap at 18h}}} \right) \, /{\text{ Area of wound gap at }}0{\text{h}}} \right] \, \times { 1}00$$

Statistical analysis

Triplicate experiments were conducted for all analyses. BBD data were processed using the rsm package in R software. Model validation was carried out via a one-sample t test to compare predicted and observed responses. One-way ANOVA with Fisher’s LSD test was used to assess differences in extraction methods, bioactive compound content, and antioxidant activities. A significance threshold of p < 0.05 was applied. Results are reported as mean ± standard deviation.

Results and discussion

Preliminary assessment of extraction parameters

Extraction solvent

Solvent selection significantly affects extraction efficiency, compound specificity, and environmental impact [27]. Figure 1 shows the TPC of hemp stems extracted with different solvents, ranging from 0.77 ± 0.05 to 1.13 ± 0.04 mg GAE/g. The highest TPC was obtained using 60% w/v aqueous propylene glycol, followed by butylene glycol, glycerol, and water. Statistically significant differences were observed among all solvents (p < 0.05). The lower TPC obtained with butylene glycol compared to propylene glycol may be explained by their polarity differences. Although both contain two hydroxyl groups, the longer hydrocarbon chain in butylene glycol decreases its polarity, thus reducing its extraction efficiency [28]. Glycerol, despite its high polarity due to three hydroxyl groups, yielded less phenolics, likely due to its high viscosity, which limits mass transfer during extraction [29]. Notably, all polyol–water solvent systems outperformed water alone, consistent with previous findings on the superior solubilizing properties of aqueous polyols [17, 30, 31]. Accordingly, the aqueous propylene glycol (60% w/v) was selected for further investigation due to its superior phenolic yield.

Fig. 1

Preliminary TPC results are shown, with distinct letters representing significant differences within each factor level (p < 0.05)

Extraction power

Ultrasonic power significantly influenced the release of phenolic compounds in this study. Extraction at 100% ultrasonic power (98 W/cm²) yielded the highest total phenolic content (1.11 ± 0.03 mg GAE/g), compared to 80% (78 W/cm²; 1.05 ± 0.01 mg GAE/g) and 50% (49 W/cm²; 1.03 ± 0.03 mg GAE/g), as shown in Fig. 1. The observed increase in TPC at higher power levels was statistically significant (p < 0.05) and is attributed to intensified acoustic cavitation effects that promote enhanced cell disruption and mass transfer [32]. Similar findings were reported by Luo et al. (2020), who demonstrated that TPC rose proportionally with ultrasonic power from 65 to 455 W [33]. These results are consistent with those of Reche et al. (2021), who noted improved yields of bioactive compounds under optimized ultrasound power density [34]. However, excessive ultrasound power may adversely affect extraction. Vo and Le (2014) also documented a decline in phenolic content and antioxidant activity in rose myrtle juice when ultrasound power exceeded 35 W/g [35]. Therefore, based on the enhanced yield without evidence of degradation, 100% power (98 W/cm²) was chosen as the optimal condition for further extraction in this study.

Extraction frequency

Ultrasonic frequency is a key factor influencing the extraction efficiency of bioactive compounds by affecting cavitation intensity and mechanical cell disruption [7]. In this study, TPC obtained at 35 kHz was significantly exceeding than at 53 kHz, as presented in Fig. 1 (p < 0.05). The enhanced extraction at 35 kHz is attributed to the generation of larger cavitation bubbles that exert stronger mechanical forces, thereby facilitating greater matrix disruption and phenolic release. In contrast, 53 kHz produces weaker cavitation, resulting in reduced efficiency [32]. Previous studies have similarly shown lower frequencies to be more effective; for instance, Dranca and Oroian (2017) reported optimal TPC extraction at 25 kHz from eggplant peel [36]. In addition, elevated frequencies have been associated with decreased TPC due to phenolic degradation caused by oxidative species such as hydrogen peroxide [37, 38]. Consequently, 35 kHz was selected as the optimal frequency for further analysis.

Extraction duration

Extraction duration in UAE significantly influences the release of bioactive compounds by enhancing cell wall disruption and solvent diffusion. However, prolonged sonication may lead to thermal degradation of thermolabile constituents [39]. In this study, the highest TPC was obtained at 45 min (1.24 ± 0.01 mg GAE/g), significantly exceeding than values obtained at 5, 15, 30, and 60 min (Fig. 1). TPC declined to its lowest at 60 min (1.08 ± 0.03 mg GAE/g), suggesting degradation due to extended cavitation-induced heat generation. Shorter durations (e.g., 5 and 15 min) may have been insufficient for effective matrix disruption, while excessive exposure at 60 min may have triggered oxidative or thermal breakdown of phenolics. These findings align with previous studies reporting decreased phenolic yields with excessive sonication [40], although other reports, such as Ez Zoubi et al. (2021), observed continuous increases in TPC with longer durations [41]. Overall, 45 min was identified as the optimal extraction duration and was selected for subsequent analyses.

Concentration of solvent

The efficiency of phenolic extraction was significantly influenced by COS, which affects polarity matching and mass transfer processes [42]. Among the tested concentrations, 40% COS yielded the highest total phenolic content (1.39 ± 0.04 mg GAE/g), significantly surpassing the remaining groups (p < 0.05), as demonstrated in Fig. 1. Conversely, the lowest TPC (0.05 ± 0.02 mg GAE/g) was observed with 100% COS. This outcome supports prior evidence that phenolic solubility and solvent–matrix interactions are optimized at intermediate polarity levels [43]. These results are consistent with the findings of Myo and Khat-Udomkiri (2022), who reported that 40% aqueous propylene glycol was the most effective solvent for extracting phenolic compounds from coffee pulp [17]. Additional studies have similarly highlighted the superior performance of aqueous solvent mixtures in extracting bioactive compounds [30, 31, 44]. Thus, 40% aqueous propylene glycol was selected as the optimal solvent system for further analysis.

Solvent–solute ratio

The SSR plays a crucial role in optimizing solvent penetration, compound solubility, and mass transfer, ultimately influencing extraction yield and the bioavailability of phenolic compounds [7]. A clear trend in TPC was observed across different SSR values (Fig. 1), with the highest TPC recorded at an SSR of 30:1 (1.26 ± 0.04 mg GAE/g). Similar values were observed at 35:1 (1.22 ± 0.20 mg GAE/g) and 40:1 (1.24 ± 0.03 mg GAE/g), with no significant differences among them (p > 0.05). Lower TPC values were associated with 25:1 (1.16 ± 0.02 mg GAE/g) and 20:1 (1.01 ± 0.02 mg GAE/g), with the latter being significantly lower than higher ratios (p < 0.05). The optimal performance at 30:1 is likely due to adequate solvent volume facilitating efficient diffusion, while further increases did not enhance extraction due to solute saturation [45]. At lower SSRs, reduced solvent availability likely impeded phenolic release. Previous studies support these findings, indicating that higher SSR enhances mass transfer and TPC up to a threshold, beyond which extraction efficiency plateaus or declines [46,47,48].

Extraction method optimization using selected experimental design

Based on the findings of initial experiments, subsequent optimization was evaluated on the enhancement of ED, SSR, and COS using BBD, with ultrasonic power and frequency maintained at 98 W/cm² and 35 kHz, respectively, as shown in Table 1. The corresponding response variables, including bioactive compound content and antioxidant activity, are summarized in Table 2. This design enabled the identification of optimal extraction parameters and provided insight into the interactive effects of the variables on the target outcomes.

Obtaining the best fit models according to the experimental design.

The statistical significance of each model term was determined based on its corresponding p value. In addition, the "lack-of-fit" F value served as a diagnostic measure to assess model adequacy, where a non-significant result is indicative of a good fit. In this study, each model for all responses is statistically significant (p < 0.05) and lack-of-fit terms were non-significant ((p ≥ 0.05). Analysis of variance (ANOVA) for the quadratic models developed to predict optimal responses revealed that all models were statistically significant (p < 0.05), as summarized in Table 3 and supplementary Table S1. The reliability of the developed models was confirmed by comparing the adjusted R² and R² values for each response variable. The adjusted R² values were 0.88 (TPC), 0.79 (TFC), 0.89 (TTC), 0.75 (DPPH), and 0.87 (FRAP), which were closely aligned with their corresponding R² values of 0.94, 0.90, 0.95, 0.88, and 0.94, respectively. These results suggest a strong fit between the experimental data and the predicted outcomes, indicating the adequacy of the response surface models. Second-order polynomial equations were derived using response surface methodology to describe the effects of three independent variables: ED (A), SSR (B), and COS (C), on each of the response variables. These equations are as follows:

$${\text{Y}}\left( {{\text{Phenolic}}} \right) \, = { 1}.0{7} + 0.0{\text{1A}} + 0.0{\text{5B}} - 0.0{\text{8C}} + 0.{\text{13AB}} + 0.0{\text{6AC}} - 0.0{\text{2BC}} - 0.{\text{13A}}^{{2}} - 0.0{\text{3B}}^{{2}} + 0.0{\text{1C}}^{{2}}$$

(2)

$${\text{Y}}\left( {{\text{Flavonoid}}} \right) \, = { 4}.{15} + 0.{\text{16A}} - 0.0{\text{6B}} + {1}.{\text{35C}} + 0.0{\text{8AB}} + 0.{\text{63AC}} + 0.{\text{15BC}} + 0.{\text{27A}}^{{2}} - 0.{\text{67B}}^{{2}} + 0.{\text{86C}}^{{2}}$$

(3)

$${\text{Y}}\left( {{\text{Tannin}}} \right) \, = { 9}.{1}0 + 0.{\text{28A}} + {1}.{\text{19B}} + {2}.0{\text{3C}} + 0.{6}0{\text{AB}} + 0.{\text{69AC}} - 0.{\text{28BC}} - 0.{\text{45A}}^{{2}} + 0.{\text{41B}}^{{2}} - {1}.{\text{34C}}^{{2}}$$

(4)

$${\text{Y}}\left( {{\text{DPPH}}} \right) \, = { 1}.{94} + 0.0{\text{5A}} + 0.{\text{44B}} + 0.{\text{16C}} + 0.0{\text{9AB}} + 0.{\text{12AC}} + 0.{\text{14BC}} + 0.{\text{34A}}^{{2}} + 0.0{\text{6B}}^{{2}} - 0.{\text{37C}}^{{2}}$$

(5)

$${\text{Y}}\left( {{\text{FRAP}}} \right) \, = { 2}.{44} + 0.0{\text{1A}} + 0.{3}0{\text{B}} + 0.0{\text{3C}} + 0.{\text{17AB}} + 0.0{\text{8AC}} + 0.{\text{14BC}} - 0.{\text{29A}}^{{2}} - 0.0{\text{7B}}^{{2}} - 0.{2}0{\text{C}}^{{2}}$$

(6)

Table 3 Model performance assessed using the lack-of-fit test, R² values, and the statistical significance (p values) of the regression coefficients for each response variable

Analysis of response surfaces

The statistical analysis revealed that the evaluated factors significantly influenced the response variables. As summarized in Table 3 and supplementary Table S2, model terms with p values below 0.05 were considered statistically significant. The interactive effects of ED, SSR, and COS on the response variables are depicted in the three-dimensional response surface plots presented in Fig. 2 and supplementary Fig. S1.

Fig. 2

Response surface plots showing the interaction effects of SSR and COS (a), ED and COS (b), and ED and SSR (c) on FRAP activity

Linear effects of extraction parameter

SSR and COS exhibited significant linear effects on the extraction of bioactive compounds and antioxidant activity. Increasing SSR significantly enhanced the recovery of all compounds except TFC (p < 0.05), likely due to improved solute diffusion at higher solvent volumes, consistent with enhanced mass transfer reported in UAE of medicinal plants [49, 50]. In contrast, COS positively influenced TFC and TTC but had a negative effect on TPC (p < 0.05), suggesting that excessive solvent polarity may interfere with hydrogen bonding, thereby reducing phenolic extraction. Similar solvent saturation effects have been documented in previous extraction processes [51].

Interaction effects of extraction parameters

The interactive effects of ED, SSR, and COS were statistically significant for key bioactive constituents and antioxidant activities. A positive interaction between ED and SSR enhanced TPC and FRAP (p < 0.05), suggesting improved phenolic diffusion with longer extraction times and increased solvent volume. Similar synergistic effects have been reported for phenolic extraction from Ajwa date and coffee pulp [17, 52]. In addition, the combination of ED and COS significantly improved both TPC and TFC (p < 0.05), likely due to enhanced solubility and mass transfer of phenolic and flavonoid compounds under higher solvent concentrations. This finding aligns with reports on Camellia sinensis flower extractions [53]. Significant ED–SSR and SSR–COS interactions also increased FRAP values, underscoring the need for optimized solvent parameters to achieve maximal antioxidant potential, consistent with trends seen in coffee pulp extractions [17].

Quadratic effects of extraction parameters

Quadratic effects revealed non-linear responses among extraction parameters. The quadratic term of ED significantly reduced TPC and FRAP values, indicating degradation of phenolics and decreased antioxidant capacity at extended durations—an effect consistent with previously reported degradation under prolonged extraction [31, 53]. Conversely, ED exhibited a significant positive effect on DPPH scavenging activity, suggesting improved release of alternative antioxidant compounds over time. This pattern corresponds with studies showing enhanced antioxidant activity despite declining phenolic content [31]. A significant negative quadratic effect of SSR on TFC was observed, highlighting a peak point beyond which TFC declined. The COS showed a positive quadratic effect on TFC but negative effects on TTC, DPPH, and FRAP, suggesting selective enhancement of flavonoid extraction and reduced efficiency for tannin and antioxidant recovery at higher COS levels. These findings align with prior evidence of solvent-selective extraction behavior [17, 53, 54].

Confirmation of the model validity

Optimization was performed to determine the ideal combination of process parameters for maximizing response outputs (Table 4). The optimal conditions identified were 30 min of ED, 28.25 mL/g SSR, and 32.72% w/v COS. As shown in Table 5, the observed experimental values under these optimized settings closely matched the model predictions, with no statistically significant differences across all responses (p ≥ 0.05). These findings confirm the adequacy and reliability of the response surface models used.

Table 4 Restriction of independent parameters and response variables for optimization

Table 5 Results from confirmation of the model validity

Comparison of UAE with conventional maceration

The comparison of extraction conditions using UAE and maceration methods with two solvent systems—aqueous propylene glycol (HUPG and HMPG) and aqueous ethanol (HUE and HME) is presented in Table 6. All extractions were conducted under a constant SSR of 28.25 mL/g and COS of 32.72%, with either 30 min (UAE) or 24 h (maceration) of ED. Among the tested conditions, the maceration method with the selected solvent (HMPG) was found to have the highest TPC at 1.22 ± 0.04 mg GAE/g sample, which was significantly exceeding than that obtained with ultrasound using ethanol (HUE; 1.06 ± 0.06 mg GAE/g, p < 0.05). However, no statistically significant differences were observed between HMPG and HUPG or HME. The HUPG condition also demonstrated a comparably high TPC value (1.13 ± 0.03 mg GAE/g), suggesting that ultrasound-assisted extraction can provide phenolic yields comparable to prolonged maceration. TFC and TTC showed no statistically significant differences across all conditions (p ≥ 0.05), indicating that both extraction methods and solvent types exerted minimal impact on these particular classes of compounds under the tested parameters. In contrast, DPPH radical scavenging activity showed clear variation among treatments. HUPG exhibited the highest antioxidant activity (2.12 ± 0.03 mg AAE/g), which was significantly greater than that of HUE (1.63 ± 0.22 mg AAE/g, p < 0.05), and also higher than values obtained from both maceration methods. Similarly, HUPG and HMPG yielded the highest FRAP values (2.06 ± 0.06 and 1.95 ± 0.10 mg AAE/g, respectively), suggesting that ultrasound in combination with aqueous propylene glycol enhances the ferric reducing antioxidant power more effectively than ethanol-based or maceration methods. These findings highlight the effectiveness of ultrasound-assisted extraction, particularly when paired with the selected solvent system, in enhancing antioxidant activity while maintaining comparable yields of polyphenolic compounds. The improved performance of UAE over conventional maceration may be attributed to the cavitation effect, which enhances cell wall disruption and facilitates solvent penetration, thereby increasing the release of antioxidant constituents within a shorter extraction time. Overall, these findings underscore the effectiveness of propylene glycol as a solvent and UAE as a rapid and efficient extraction method. Previous research reinforced the advantages of polyol-based green extraction and the superior performance of ultrasound-assisted techniques in maximizing bioactive compound recovery. In previous studies of phenolic extraction of tea leaves, coffee pulp, and Eclipta alba stem, polyol-based extraction achieved the highest total phenolic, flavonoid, and tannin contents and superior cellular antioxidant activity compared to ethanol-based extraction [17, 18, 44].

Table 6 Results of comparison of HUPG, HUE, HMPG, and HME

Chromatographic profiling of phytochemicals of hemp stem extracts

UHPLC–ESI–QTOF–MS/MS was employed in negative ion mode to profile phenolic compounds in hemp stem extracts, with compound identification confirmed by MS/MS fragmentation patterns and molecular formula generation (MFG) scores. Phytochemical constituents with MFG scores > 80 and mass errors within ± 5 ppm were reported in this study as presented in Table 7. The LC–MS/MS analysis revealed a broad spectrum of phytochemicals in both HUPG and HUE extracts, encompassing multiple bioactive classes with pharmacological activities (Figs. 3 and 4). Phenolic compounds were the most abundant group detected, including isofeuric acid, m-coumaric acid, and isobutyl salicylate. These phenolic constituents are recognized for their potent antioxidant and anti-inflammatory activities, primarily due to their ability to neutralize free radicals and mitigate oxidative stress, corroborating their presence in numerous medicinal plants traditionally employed for therapeutic purposes [55]. Terpenoids, and ubiquinone, other major classes of bioactives, were represented by compounds, such as geranyl acetoacetate, dehydrocurdione, warburganal, and myrsinone. While these compounds were detected in both HUPG and HUE, heliniorbiasone was exclusive to HUPG, and sonchuionoside C was exclusive to HUE further contributing to their distinct chemical profiles. Terpenoids are widely recognized for their diverse pharmacological activities, including anti-inflammatory, anticancer, and antimicrobial effects [56, 57]. In addition, chelidonic acid, a quinone with established antioxidant and antimicrobial properties, and paratocarpin E, an isoprenoid, were identified in HUPG extract, further emphasizing the bioactive diversity of hemp stems [58]. However, cannabidiol (CBD) and related cannabinoids are not detected in the LC–MS/MS analysis of hemp stem extracts. Cannabinoids are predominantly synthesized and stored in the glandular trichomes found on the flowers and leaves of C. sativa, while stems lack these structures, leading to minimal or no cannabinoid production in this part of the plant [59]. In addition, the biosynthesis of cannabinoids is regulated by enzymes, such as cannabidiolic acid (CBDA) synthase, which exhibit tissue-specific expression, being active primarily in the floral tissues and largely absent in stems [60]. Moreover, previous studies analyzing non-floral parts of C. sativa, such as stems and roots, consistently report an absence or negligible presence of cannabinoids [61, 62]. Notably, this study represents the first report on the chemical profiling of hemp stems using an aqueous propylene glycol solvent system and LC–MS/MS analysis, providing novel insights into the phytochemical diversity of hemp stems.

Table 7 Chromatographic profiling of phytochemicals in hemp stem extracts

Fig. 3

Mass spectrometry analysis conducted in negative ion mode presenting the total ion chromatograms of HUPG (a) and HUE (b)

Fig. 4

MS/MS spectra with distinct fragmentation patterns for isoamyl p-anisate, warburganal, and isobutyl salicylate, as shown in panels (a), (b), and (c), respectively

Cytotoxicity and cellular antioxidant activity of hemp stem extracts

Cell viability assessment in HaCaT cells

Cytotoxicity profiling in HaCaT cells revealed concentration-dependent responses to ascorbic acid (AA), allantoin, and hemp stem extracts (HUPG and HUE), reflecting their differential impacts on cell viability (Fig. 5). AA exhibited a biphasic concentration–response pattern, significantly reducing cell viability from 95.95 ± 7.87% at 0.001 mg/mL to 15.15 ± 2.07% at 1 mg/mL. While over 80% viability was retained at concentrations ≤ 0.01 mg/mL, a marked cytotoxic effect emerged at higher doses. This concentration-dependent shift reflects the dual redox nature of AA, functioning as an antioxidant at low concentrations and as a pro-oxidant at higher levels via ROS-induced cytotoxic mechanisms [63]. The tested concentrations of allantoin showed consistently high levels of cell viability, indicating its low cytotoxic potential. At 100 µg/mL and 200 µg/mL, viability remained above 94%, reaching 108.82 ± 3.47% at 200 µg/mL, while a slight reduction to 77.34 ± 6.21% was observed at 400 µg/mL. These outcomes corroborate prior findings that attribute allantoin’s biological safety to its well-documented cytoprotective and anti-inflammatory properties, which contribute to enhanced cellular proliferation and tissue repair even at relatively high doses [64]. The hemp stem extracts, HUPG and HUE, demonstrated high levels of cell viability at low concentrations, with values of 96.48 ± 2.50% and 95.75 ± 3.37%, respectively, at 1 mg/mL. However, an obvious decline in cell viability was observed at higher tested concentrations, reaching 36.43 ± 2.53% for HUPG and 41.73 ± 1.68% for HUE at 50 mg/mL, indicating concentration-dependent cytotoxicity. These findings suggest that although the extracts are well-tolerated at lower doses, their bioactive profiles—particularly rich in phenolics and terpenoids—may contribute to cytotoxic effects when administered at higher concentrations. This pattern aligns with previous studies reporting that polyphenols can exert dual effects: antioxidant benefits at low doses and potential pro-oxidant or cytotoxic actions at elevated concentrations [65].

Fig. 5

Treatment with AA, allantoin, HUPG, and HUE exhibited varying cytotoxicity in HaCaT cells, with significant differences from the control group marked by * (p < 0.05)

Cell viability assessment in RAW 264.7 macrophage cells

As shown in Fig. 6, both HUPG and HUE exhibited low cytotoxic effects on HaCaT cells, as evidenced by their high cell viability percentages. HUPG preserved over 80% viability even at the highest concentration tested (50 mg/mL), with a range from 99.65 ± 0.65% to 80.90 ± 1.71%. In parallel, HUE maintained cell viability between 106.08 ± 0.27% and 89.60 ± 0.45%, demonstrating comparable biocompatibility to that of the reference anti-inflammatory agent, indomethacin. Indomethacin, used as a reference anti-inflammatory agent, showed non-cytotoxic effects, with viability decreasing from 103.67 ± 2.11% at 0.003125 mg/mL to 84.18 ± 1.41% at 0.1 mg/mL. These results suggest that all tested samples were well-tolerated by cells at the evaluated concentrations. The findings are consistent with previous reports on Tetragonula sapiens propolis, where high cell viability (75.99 ± 2.44%) was maintained even at 200 μg/mL [66].

Fig. 6

Effects of indomethacin (I), HUPG, and HUE on cell viability of RAW 264.7 macrophages. Significant difference from the control group is denoted by ‘*’ (p < 0.05)

The antioxidant potential of hemp stem extracts

To assess antioxidant activity, HaCaT cells were exposed to 300 µM H₂O₂ to induce oxidative stress, resulting in a significant reduction in cell viability (62.45 ± 1.97%) as presented in Fig. 7. The cytoprotective effects of hemp stem extracts (HUPG and HUE) and ascorbic acid (AA) were evaluated under these conditions. AA exhibited a concentration-dependent protective response, increasing cell viability to 69.72 ± 5.87% at 0.001 mg/mL and 73.55 ± 1.38% at 0.01 mg/mL. These findings are consistent with AA’s established capacity to neutralize reactive oxygen species (ROS) and attenuate oxidative injury in keratinocytes [63]. The findings revealed a concentration-dependent increase in cell viability following treatment with hemp stem extracts. Specifically, HUPG extract at 1 mg/mL enhanced the viability of the cells to 66.28 ± 2.59%, with a notable increase to 75.29 ± 2.21% at 5 mg/mL, indicating strong antioxidant potential. HUE extract exhibited a comparable, though slightly less pronounced, effect, with cell viability rising from 66.02 ± 6.36% to 74.71 ± 3.83% over the same concentration range. These protective effects against oxidative stress can be attributed to the presence of phenolic compounds identified through LC–QTOF–MS/MS analysis (Table 7), including isofeuric acid and acrovestone, both of which are well-documented for their ability to neutralize reactive oxygen species and reduce cellular damage [67, 68]. In addition, the presence of m-coumaric acid, unique to HUPG, may explain its superior antioxidant activity, as it has been reported to exhibit strong ROS-neutralizing effects [69]. The terpenoids identified in the extracts, such as dehydrocurdione and heliniorbiasone, further contribute to the observed antioxidant activity as terpenoids are widely recognized for their ability to enhance cellular defenses against oxidative stress by modulating antioxidant enzyme activities and directly scavenging ROS [56]. These findings highlight the potential of hemp stem extracts, particularly HUPG, as a promising source of natural antioxidants for mitigating oxidative stress in dermal applications.

Fig. 7

Cellular antioxidant effects of AA, HUPG, and HUE under hydrogen peroxide-induced stress are presented, with * indicating significance versus the control and # indicating significance versus the H₂O₂ group (p < 0.05)

Evaluation of antimicrobial potential

The antimicrobial activity of hemp stem extracts prepared by different extraction methods was assessed against S. aureus (SA), S. epidermidis (SE), E. coli (EC), and P. aeruginosa (PA)—pathogens commonly associated with skin infections. The activities of HUPG and HUE extracts were compared to conventional antibiotics. As shown in Table 8, the HUPG extract demonstrated inhibitory activity against all tested strains, with particularly notable effectiveness against gram-negative bacteria. Among these, P. aeruginosa exhibited the highest sensitivity, with  MIC and MBC values of 12.5% v/v and 50% v/v, respectively. In contrast, gram-positive bacteria displayed comparatively lower susceptibility to the extract. The greater antibacterial potential of HUPG compared to HUE may be explained by the presence of bioactive constituents exclusive to the HUPG extract, including paratocarpin E, chelidonic acid, and helinorbisabone—compounds recognized in the literature for their potent antimicrobial effects. [70,71,72]. In addition, previous studies have reported the antibacterial properties of hemp stems and their use in antibacterial finishing agents and surgical devices [59, 73, 74]. More recent findings have demonstrated the antibacterial activity of hemp hurd powder against E. coli [75]. Hemp stem extracts, particularly HUPG, demonstrated antimicrobial activity, though their efficacy was lower than that of standard antibiotics, such as ampicillin and ciprofloxacin. These results suggest that while hemp extracts show promise as natural antimicrobials, their clinical application may require higher concentrations or further optimization, such as active compound isolation or combination strategies. This study provides preliminary evidence supporting their potential against common skin pathogens.

Table 8 Results of anti-bacterial potential of plant extracts

Anti-inflammatory potential evaluation of plant extracts

In vitro evaluation of anti-inflammatory property

The comparative analysis of anti-inflammatory property of hemp stem extracts revealed that HUPG exerts more potent anti-inflammatory effects than HUE, as evidenced by both 5-LOX inhibition and NO radical scavenging activities as presented in Table 9. At 31.25 mg/mL, HUPG significantly inhibited 5-LOX activity (22.53 ± 5.10%), nearly double the inhibition observed with HUE (11.22 ± 1.38%). Furthermore, HUPG exhibited enhanced NO scavenging capacity (1.41 ± 0.38 mg GAE/g) relative to HUE (1.25 ± 0.21 mg GAE/g), with statistically significant differences (p < 0.05). These results underscore the potential role of hydroxycinnamic acid compounds, particularly isoferulic acid, in mediating the anti-inflammatory properties of HUPG through mechanisms involving lipoxygenase pathway modulation [76]. In addition, a previous study reported that Viola odorata extract containing m-coumaric acid and other bioactive compounds demonstrated downstream effect on prostaglandin E-2 (PGE-2) and NO [77]. This could contribute to the higher anti-inflammatory activity of HUPG compared to HUE. In addition, the terpenoids identified in extracts, such as dehydrocurdione and heliniorbiasone, also contribute to its enhanced anti-inflammatory activity as terpenoids are known to downregulate pro-inflammatory cytokines and inhibit the activity of inflammatory enzymes, such as 5-LOX and cyclooxygenase (COX), thereby reducing inflammation [78]. Although HUPG demonstrated promising anti-inflammatory activity, its efficacy remains lower than that of NDGA, a potent 5-LOX inhibitor (IC₅₀ = 1.26 ± 0.11 µg/mL). Nevertheless, the broad range of bioactive compounds in HUPG provides a multifaceted approach to inflammation management, suggesting its potential as a complementary or alternative therapeutic agent. Future studies should focus on isolating individual bioactive compound and investigating their specific mechanisms of action to better understand their contributions to the overall anti-inflammatory activity of hemp stem extracts.

Table 9 Evaluation of anti-inflammatory properties of hemp stem extracts

Evaluation of nitric oxide inhibitory activity in LPS-stimulated raw 264.7 macrophages

This study investigated the anti-inflammatory properties of hemp stem extracts by evaluating their ability to inhibit NO production in RAW 264.7 macrophages. Using indomethacin as a positive control, both HUPG and HUE extracts demonstrated concentration-dependent NO inhibition, as shown in Fig. 8. A concentration-dependent suppression of NO production was observed with HUPG, ranging from 2.05 ± 0.14% at 1 mg/mL to a maximum of 46.78 ± 1.79% at 50 mg/mL. HUE showed a similar inhibitory trend, achieving 51.03 ± 2.02% at the highest tested dose, though it was less effective at lower concentrations. Indomethacin, the standard control, exhibited the strongest effect with an IC₅₀ of 0.015 mg/mL. Overall, the results support the anti-inflammatory potential of hemp stem extracts, with HUPG displaying greater effectiveness at lower concentrations. The findings align with other in vitro assays, such as 5-LOX and NO inhibition tests. The activity observed is likely associated with phenolic compounds present in the extracts, including isoferulic acid and m-coumaric acid, which have been previously shown to inhibit pro-inflammatory pathways, such as NF-κB and MAPK, resulting in reduced production of cytokines, such as TNF-α, IL-6, and IL-1β [77]. A previous study reported that chelidonic acid supresses the inflammatory mediators, including cytokines and the expression of COX-2 and iNOS in vivo and in vitro [79]. The extracts also contain terpenoids, alkaloids and ubiquinones, which have been implicated in modulating NF-κB and JAK–STAT pathways, leading to the suppression of inflammatory mediators [80, 81]. Despite its promising effects, the potency of hemp extracts remains lower than that of indomethacin, with an IC₅₀ of 0.015 mg/mL. However, the findings suggest that hemp stem extracts, particularly HUPG, exhibit significant anti-inflammatory properties supporting their potential use as complementary therapeutic agents in managing inflammation-related disorders. Future research should focus on isolating individual bioactive compounds and elucidating their precise mechanisms in preclinical and clinical models.

Fig. 8

Nitric oxide inhibitory effects of indomethacin, HUPG, and HUE on RAW 264.7 macrophages. p < 0.05 vs. control (*)

Wound healing via cell scratch assay

The results of the cell scratch assay indicate that hemp stem extracts exert differential effects on keratinocyte migration depending on their composition and concentration, as presented in Fig. 9. Notably, HUPG demonstrated a concentration-dependent increase in wound closure, reaching 44.62 ± 4.41% at 1 mg/mL and further improving to 51.92 ± 6.05% at 5 mg/mL. These values surpassed both the untreated control (27.69 ± 4.23%) and the reference compound, allantoin (43.72 ± 3.30% at 200 µg/mL). In comparison, HUE exhibited modest activity, with wound closure rates of 36.38 ± 2.45% and 39.75 ± 2.91% at 1 and 5 mg/mL, respectively. These findings parallel the trends observed in antioxidant and anti-inflammatory assays, suggesting a strong correlation between bioactivity and the phytochemical richness of the extracts. HUPG's enhanced efficacy may be attributed to its broader spectrum of bioactive constituents that synergistically promote tissue repair, while the lower activity of HUE could reflect the absence or lower concentrations of such components. These observations are supported by existing literature that highlights the role of hemp-derived compounds in modulating inflammatory responses, mitigating oxidative stress, and enhancing regenerative pathways essential for wound healing [82, 83]. A previous study reported that C. sativa L. extracts and CBD inhibit pro-inflammatory mediators such as IL-8, VEGF, and MMP-9 in keratinocytes and fibroblasts by modulating the NF-κB pathway, thereby controlling excessive inflammation and promoting extracellular matrix remodeling [82]. Similarly, a previous study found that hemp water extracts rich in CBDA, CBD, and rutin significantly attenuate oxidative stress and apoptosis, while regulating IL-6 and PGE2 production, both of which are critical for tissue regeneration [84]. In this study, the variation in solvent systems and plant parts used highlights the need for further investigation into the mechanism of action underlying the wound-healing efficacy of aqueous propylene glycol extracts.

Fig. 9

a Scratch assay assessed keratinocyte migration across treatment groups over an 18-h period. Images at 0 and 18 h show the extent of wound closure under control, at the concentration of 200 μg/mL of allantoin, HUPG (1 and 5 mg/mL), and HUE (1 and 5 mg/mL) conditions (10 × magnification). Scratch area reduction reflects progressive cell migration. b Quantitative analysis of wound closure revealed significant differences among treatments. Asterisks (*) indicate statistical significance versus the control, and hash symbols (#) denote significance compared to the allantoin group (p < 0.05)

Limitations of current study

This study highlights the wound-healing potential of hemp stem extracts; however, certain limitations should be noted. The main constraint of extracting phenolic compounds from hemp stems with propylene glycol is its low volatility and high boiling point, which makes solvent concentration and recovery difficult compared to conventional alcohols. However, this drawback can be offset by the fact that propylene glycol is a safe, pharmaceutically accepted excipient that allows the extract to be directly incorporated into topical or dermocosmetic formulations without the need for solvent removal, thereby preserving thermolabile phenolics. Pre-treatments such as drying, milling, and enzymatic cell wall disruption may further enhance extraction efficiency, while comprehensive biological evaluations (antioxidant, anti-inflammatory, antimicrobial, and cytotoxicity assays) are essential to confirm the practical applicability of the extract. In addition, the lack of quantitative profiling of bioactive compounds prevents clear attribution of activity to specific constituents, despite the promising performance of HUPG. Incorporating techniques such as HPLC or LC–MS/MS in future research will enable accurate measurement of active components and clarify their roles in regeneration. In addition, given the in vitro nature of this study, the findings may not entirely reflect in vivo conditions. Therefore, further in vivo investigations are essential to validate these results, optimize dosing strategies, and assess clinical safety and efficacy.

Conclusion

This study optimized ultrasound-assisted extraction of hemp stem using propylene glycol, revealing its strong potential for wound healing and related therapeutic applications. Optimal extraction conditions—30 min, 28.25 mL/g liquid-to-solid ratio, and 32.72% solvent concentration—were identified via Box–Behnken design and response surface methodology. The resulting HUPG extract showed higher yields of total phenolics, flavonoids, and tannins, along with enhanced antioxidant activities (DPPH and FRAP), outperforming conventional extraction methods. Chromatographic phytochemical analysis confirmed the presence of bioactive compounds, such as isofeuric acid, m-coumaric acid, and isobutyl salicylate, supporting the extract's pharmacological value. HUPG demonstrated in vitro antioxidant, antibacterial, anti-inflammatory, and wound-healing effects, suggesting its potentials to be used in cosmetic and pharmaceutical formulations. Further in vivo and clinical validation, along with quantitative analysis of key active compounds, is recommended. These findings highlight hemp stem as a promising, eco-friendly resource for sustainable product development in healthcare and skincare.