Impact of Plant Growth Regulators on Callus Induction in Cannabis sativa L.

Abstract

Callogenesis is a fundamental step in plant biotechnology and tissue culture, providing the basis for multiple scientific and practical applications. In this study, the impact on callogenesis of different plant growth regulators was studied on Cannabis sativa L. (a non-commercial genotype of hemp), with the objective of identifying the most suitable combination for the establishment of vigorously growing, friable calli. Forty-nine media combinations were evaluated using four PGRs: two auxins (2,4-dichlorophenoxyacetic acid, naphthaleneacetic acid) and two cytokinins (6-benzylaminopurine, kinetin). Parameters such as percentage of callus induction, proliferation, colour, texture, and growth area were assessed. Three media were identified for further spectrophotometric assays and targeted gene expression analysis: the first containing 2,4-dichlorophenoxyacetic acid 1.5 µM and benzylaminopurine 1.5 µM, the second with 2,4-dichlorophenoxyacetic acid 1.5 µM and kinetin 1.5 µM and the third supplemented with 2,4-dichlorophenoxyacetic acid 4.5 µM and kinetin 1.5 µM. The last medium proved to be superior in terms of vigour, friability and phenolic content and showed increased expression of genes involved in the early steps of the phenylpropanoid pathway. These findings highlight the central role of auxin–cytokinin interactions in regulating both callus formation and secondary metabolism. The optimised medium opens the way to subsequent biotechnological applications relying on the cultivation of plant cell suspension cultures.

1. Introduction

Cannabis sativa L. is a highly versatile plant valued for its fibres, nutrient-rich seeds [1] and, most notably, its diverse array of bioactive secondary metabolites including cannabinoids, flavonoids, and terpenes [2,3]. These compounds highlight the plant’s pharmaceutical, cosmetic and nutraceutical relevance, fuelling growing interest in sustainable and scalable biotechnological production systems.

In this context, plant tissue culture techniques have re-emerged as powerful tools: after experiencing a boom of research activity in the 1990s, followed by a period of relative decline, tissue culture has recently regained attention due to its immense biotechnological potential [4]. Beyond clonal propagation and germplasm conservation [5], in vitro culture systems enable controlled studies of plant development and offer alternative routes to produce valuable metabolites [6,7,8,9]. This renewed focus positions tissue culture methods at the intersection of fundamental research and applied biotechnology, particularly for multi-purpose species like C. sativa.

At the core of many tissue culture protocols lies callus induction (a.k.a. callogenesis), a process that demonstrates the capacity of plant tissues to dedifferentiate, i.e., lose their committed fate and re-enter the cell cycle: this step is often found in experimental strategies aimed at regeneration, transformation and establishment of suspension cultures. Callogenesis refers to the induction and proliferation of a callus, an unorganised mass of rapidly dividing cells, typically induced by mechanical stress (e.g., wounding) and controlled by plant growth regulators (PGRs) [10]. Several factors, among which are nutrients and organic components, such as PGRs and sucrose concentration, influence the processes of organogenesis, somatic embryogenesis and callogenesis [11]. In plant tissue culture, PGRs play a key role in regulating important processes, such as cell division, differentiation, and organogenesis through their antagonistic and synergistic interactions [12].

Concerning callogenesis, these factors also influence key traits such as texture (friability/compactness) and colour (e.g., white, cream, brown, pigmented), which are often correlated with the morphological and regenerative characteristics of calli [13], as well as their metabolic activity [14]. With respect to callus formation, auxins and cytokinins are the most widely used PGRs. Their balance influences whether tissues undergo differentiation or dedifferentiation [10]. Notably, callus induction is in general favoured by an intermediate auxin/cytokinin ratio [10]. However, optimisations are necessary depending on the species.

Once vigorously growing, friable calli are established, cell suspension cultures can be set up by transferring them to liquid culture media under shaking. This step is crucial and limiting for the subsequent optimisation of bioprocesses in bioreactors. Because of this central role, robust and effective callus induction is a prerequisite for any advanced plant biotechnology platform.

In the case of C. sativa, optimised callogenesis protocols are not only essential for plant regeneration [15,16] and conservation of elite genotypes, but also for advanced applications such as CRISPR/Cas9 gene editing [17], elicitor-based enhancement of metabolite production [18], and cell suspension culture development [19]. In this respect, optimising callus induction systems in C. sativa represents a strategic step towards harnessing the biotechnological potential of this plant, as callus cultures are essential for scalable secondary metabolite production [20]. It is indeed well-known that C. sativa produces a wide range of secondary metabolites, including phenolic compounds [21]. Among phenolics, flavonoids constitute a major subclass characterised by structural diversity and are particularly interesting in the light of their bioactivities. C. sativa callus cultures thus provide a controlled and versatile system to investigate phenolic biosynthesis and accumulation, supporting fundamental studies on the regulation of secondary metabolism under defined hormonal and environmental conditions, as well as applied research aimed at optimising metabolite production for biotechnological applications.

In the present study, the effects of four widely used PGRs, i.e., 2,4-dichlorophenoxyacetic acid (2,4-D), naphthalene acetic acid (NAA), 6-benzylaminopurine (BAP) and kinetin, were studied on Murashige and Skoog (MS) medium, which was selected due to its well-balanced macronutrient and micronutrient composition [22], to find the optimal conditions for callogenesis in C. sativa. This multi-media screening study provides an optimised callogenesis protocol for a non-commercial genotype, Bolonska, sourced from the N.I. Vavilov Institute of Plant Genetic Resources (VIR). The choice of this non-commercial genotype was due to its accessibility through a germplasm collection, its prior characterisation for phytochemical composition [23] and its availability in vitro in sufficient quantities, enabling controlled and repeated experimentation.

2. Materials and Methods

2.1. Plant Material and Preparation of Explants

Micropropagation was carried out as previously described [23]. Briefly, seeds of the genotype Bolonska (VIR; k-6) were sown in pots containing a 1:2 mixture of sand and potting soil and grown for 8 weeks in growth chambers (Fitotron, Weiss Technik, Liedekerke, Belgium) under controlled conditions (16 h light/8 h dark, 25/20 °C day/night, 60% humidity). Plants were then harvested and used as explant sources: stems were cut into segments of ca. 5 cm with one internode, surface-sterilised and rinsed 3 times with sterile water. Explants were subsequently cultured in Magenta GA-7 boxes (Merck KGaA, Darmstadt, Germany), containing solid MS medium [22] supplemented with MS vitamins, 3% (w/v) sucrose, and 0.8% (w/v) agar (pH 5.8), and maintained under sterile conditions with a 16 h/8 h photoperiod at 22 °C. The plants were micropropagated for 8 weeks on MS medium without hormones to generate enough material for callogenesis optimisation.

2.2. Callogenesis Optimisation

Callogenesis was carried out on MS medium supplemented with MS vitamins, 3% (w/v) sucrose, 0.8% (w/v) agar, adjusted to pH 5.8 before autoclaving, and different combinations of NAA, BAP, kinetin and 2,4-D. The different combinations of PGRs used are described in

Table 1. The combination of plant growth regulators (PGRs) used for the multi-media screening for callogenesis is represented as follows: yellow indicates NAA concentrations, green refers to BAP concentrations, blue indicates 2,4-D concentrations, pink shows kinetin (KIN) concentrations, and orange denotes the medium numbers. Grey-shaded boxes refer to PGR combinations already shown in the table. The arrows indicate the concentration of each PGR (NAA: yellow, BAP: green; 2,4-D: blue; KIN: pink).

NAA (µM) ↓	0.00	1.50	4.50	13.50	0.00	1.50	4.50	13.50	← BAP (µM)
0.00	1	2	3	4					0.00
5.00	5	6	7	8		26	27	28	1.50
15.00	9	10	11	12		29	30	31	4.50
45.00	13	14	15	16		32	33	34	13.50
0.00						35	36	37	0.00
5.00		17	18	19	38	39	40	41	1.50
15.00		20	21	22	42	43	44	45	4.50
45.00		23	24	25	46	47	48	49	13.50
KIN (µM) →	0.00	1.50	4.50	13.50	0.00	1.50	4.50	13.50	↑ 2,4-D (µM)

. The concentrations were determined according to the literature data [16,24,25,26,27,28,29]. The media were autoclaved (Systec 1095 VX-150 autoclave, Systec GmbH & Co. KG, Linden, Germany) before use, with a sterilisation cycle of 20 min at 121 °C. All the plant growth regulators (PGRs) were added before autoclaving.

Stem explants of 0.8–1.2 cm were excised between two nodes under sterile conditions using a scalpel and randomised before transferring them to the Petri dishes containing the different media (55 × 14 mm). Randomisation was performed to avoid biases due to putting explants from the same donor plant together on the same plate. Six Petri dishes were used per medium (with each dish containing 3 stem explants, hence 18 stem explants per treatment). The plates were placed in darkness for 4 weeks at 22 °C, and their position was randomised every 3–4 days. After 4 weeks, different parameters were scored: % of callogenesis (inferred from the total number of stems where calli were visible), extent of callus formation (one-sided, two-sided or homogeneous callus formation across the whole explant), callus proliferation (low, medium, high), colour (cream-coloured, brown or white) and texture (friable or compact). For each variable, the values were scaled to a dimensionless score between 0 and 1, where 1 represents the optimal condition and 0 represents the least optimal condition. Intermediate values were linearly scaled between these two extremes, allowing for the data to be directly compared. This normalisation approach was followed by a visual representation of the data in the form of a hierarchical clustering of the heatmaps, which was done with Cluster 3.0 [30] and rendered with Java Treeview 3.0 [31] (available at http://jtreeview.sourceforge.net/, accessed on 15 October 2025). For heatmap visualisation, normalised values (scaled between 0 and 1) were represented using a continuous colour gradient, where 0 corresponds to the least optimal condition and 1 to the optimal condition.

Subsequently, on a set of seven media showing the best scores, calli were separated from the original explants and sub-cultured on the corresponding media to calculate their growth over a time-window of 2 months. The growth area was calculated by taking photographs every 7 days and delineating the contour of each callus with ImageJ (version 1.54d) to select the best media. Three media were identified, and calli growth was again monitored after cultivation over a period of 2 months, corresponding to 2 consecutive rounds of sub-culture of 1 month each. For each of the 3 media, the calli texture, colour, total phenolic content, antioxidant activities and gene expression were evaluated.

2.3. Spectrophotometric Assays for the Determination of the Antioxidant Activity and Total Phenolic Content

Calli from each of the 3 final media were flash-frozen in liquid nitrogen, ground to a fine powder using a mortar and pestle and lyophilised for 3 days. The samples were then weighed, and an ethanol/water (80/20 v/v) mixture was finally added at a 1:30 ratio (w/v) to perform extraction. The mixture was vortexed for 5 s and sonicated for 10 min at 37 Hz, with 100% power at room temperature. The samples were then agitated for 30 min at 6.5 Hz at room temperature with the Mixer Mill MM 400 (Retsch, Veder Scientific, Haan, Germany) and centrifuged for 20 min at full speed (14,000 rpm) at 4 °C. The supernatants were transferred to 2 mL tubes (Eppendorf). The amount of total phenolic content was determined using the Folin–Ciocâlteu method [32,33]. A calibration curve was prepared with a concentration of gallic acid ranging from 0 to 0.05 mg/mL. The total phenolic content was expressed as gallic acid equivalent (GAE)/g DW (dry weight).

The antioxidant activities were determined using the Ferric Reducing Antioxidant Power (FRAP) assay [34]. In a 96-well plate, 10 µL of extract and 190 µL of FRAP reagent were pipetted. The reagent was composed of acetate buffer, ferric 2,4,6-tripyridyl-s-triazine (TPTZ) solution, and FeCl₃ at a ratio of 10:1:1 v/v/v. Three technical replicates were prepared per sample. The plate was then incubated in darkness for 20 min and the absorbance was read with a spectrophotometer (Spark20M, TECAN, Männedorf, Switzerland) at 593 nm. A standard curve with FeSO₄ was prepared (0–1 mM). The final antioxidant power was expressed as mM FeSO₄/g dry weight (DW).

2.4. RNA Extraction and Targeted Gene Expression Analysis

Calli (1-month-old) grown on the 3 selected media were harvested for targeted gene expression analysis. Three biological replicates were prepared. Total RNA was extracted from the finely ground calli using the Qiagen RNeasy Plant Mini kit (Qiagen N.V, Venlo, Netherland) with the on-column DNase treatment. The purified RNAs were quantified using a NanoPhotometer NP80 (Implen, Munich, Germany) with RNase-free water as a blank. RNA integrities were then checked by capillary gel electrophoresis on a 2100 Bioanalyzer (Agilent, Santa Clara, California, USA), using the RNA 6000 nano chip (Agilent). All RINs were >7.

One microgram of RNA was retrotranscribed to cDNA using the ProtoScript II RTase (NEB) and random hexamers. RT-qPCR was carried out in 384-well microplates prepared with a pipetting robot (epMotion, Eppendorf, Hamburg, Germany) and run using the Takyon Rox SYBR MasterMix dTTP blueMix (Kaneka Eurogentec S.A., Seraing, Belgium) in a QuantStudio™ A28569 thermal cycler (Thermo Fisher Scientific, Waltham, Massachusetts, USA). The expression was calculated using 2 reference genes (Fbox and eTIF4E), which geNORM, implemented in qBASE⁺ [35], identified as sufficient for data normalisation. The primers were previously reported [23,36,37] (Supplementary File, folder “qPCR primers”).

2.5. Cultivation of Other C. sativa Genotypes on the 3 Selected Media

Calli derived from stem explants of other genotypes propagated in vitro, namely Local 14 and Fédora 17, were initially induced and cultivated on a medium supplemented with 2,4-D 4.5 µM and kinetin 0.93 µM. Subsequently, the calli were transferred to the 3 selected media and sub-cultured twice (corresponding to a period of 2 months) to allow acclimatisation to the new culture conditions.

2.6. Statistics

Statistics were performed with the SPSS software (IBM SPSS Statistics v20) after log10-transformation of the data. Normality and homogeneity were checked with a Shapiro–Wilk and Levene’s test, respectively. Normal and homogeneous data were analysed with a one-way ANOVA followed by Tukey’s post hoc test, while non-parametric data were analysed with a Kruskal–Wallis test coupled to a Dunn’s post hoc test.

3. Results

3.1. Effect of PGRs on Callogenesis

The most effective callogenesis media were identified by assessing five key parameters: percentage of callus induction, extent of callus formation across the stem, callus proliferation, colour, and texture (Figure 1; Supplementary File, folders “Parameters scored” and “Means of the parameters”). The optimal characteristics for these parameters were defined as follows: 100% callus induction, callus formation across the entire stem segment, high proliferation of callus biomass, cream-coloured, and friable texture.

The hierarchical clustering of the normalised data obtained for each of the five parameters (Supplementary File, folders “Normalised data” and “Means of the normalised data”) revealed three main clusters when using a Pearson correlation coefficient threshold ≥0.91 (Figure 2), indicating strong similarity among the best-performing media. Cluster 1 comprises media 4, 24, 17 and 46, which did not show optimal parameters, as evidenced by a low callus proliferation. Similarly, Cluster 2, comprising media 2, 37, 3, 48, 35, 36, 44 and 49, was characterised by generally poor scores across all media, including callus proliferation, coverage, colour, and texture. Cluster 3 also resulted in non-optimal characteristics, particularly in terms of callus proliferation, colour, and texture, thereby limiting its suitability for the subsequent establishment of cell suspensions. In contrast, the sub-group within Cluster 1, including eight media formulations, i.e., media 20, 26, 13, 43, 39, 38, 42 and 47 (Table 1), showed the best results. However, medium 20 was excluded due to its suboptimal callogenesis percentage. The remaining seven media from Cluster 1, i.e., media 26, 13, 43, 39, 38, 42 and 47, were selected for further analysis to identify the most effective formulations among them. These seven media consistently induced callogenesis with high proliferation and extent of explant coverage. For simplicity, these seven media will be referred to as media I-VII (

Table 2. Parameters observed on the 7 selected media for the calli of the genotype Bolonska after a subculture lasting 65 days for the second step of the screening. KIN: kinetin. Scale bars = 0.5 cm.

Medium	Number	Original Number	Texture	Colour	Picture
NAA 45 µM	I	13	Compact	Brown
2,4-D 1.5 µM	II	38	Compact	Brown
2,4-D 1.5 µM with + KIN 1.5 µM	III	39	Friable	Creamy
2,4-D 4.5 µM with + KIN 1.5 µM	IV	43	Friable	Creamy
2,4-D 13.5 µM with + KIN 1.5 µM	V	47	Compact	Brownish
2,4-D 1.5 µM with + BAP 1.5 µM	VI	26	Friable	Creamy
2,4-D 4.5 µM	VII	42	Compact	Brownish

).

3.2. Efficiency of Selected Media in Callus Propagation and Maintenance

Calli derived from the seven selected media were sub-cultured onto their respective media for a period of 2 months. At the end of this period, callus growth, texture, and colour were again assessed. Growth was evaluated by measuring the surface area (in cm²) of each callus at day 0 (D0) and day 65 (D65) using the ImageJ software version 1.54d (Supplementary File, folder “Growth area”). At day 65 of the experiment, the growth of the calli did not differ significantly across the seven different media (Supplementary File, folder “Growth area”).

Consequently, media selection was based on qualitative parameters: callus colour and texture. Based on the previous evaluation criteria, three media were selected for further analysis: 2,4-D 4.5 µM + kinetin 1.5 µM (henceforth referred to as medium A, a.k.a. medium 43), 2,4-D 1.5 µM + kinetin 1.5 µM (medium B, a.k.a. medium 39), and 2,4-D 1.5 µM + BAP 1.5 µM (medium C, a.k.a. medium 26). Calli maintained on these three media were predominantly cream-coloured and friable (Table 2).

The calli were transferred onto fresh medium and sub-cultured for two successive cycles, each lasting 1 month (Figure 3). In Figure 3, it is possible to observe that medium B promoted better growth compared to medium A and C. Concerning colour and texture, no major differences could be observed.

3.3. Spectrophotometric Assays in the Calli Propagated on the Three Best Media

The total phenolic content and antioxidant capacity were measured to evaluate and compare the three selected media. The results are shown in Figure 4. The data indicated that media A and B had statistically significantly higher total phenolic contents than medium C (Figure 4A). With respect to the antioxidant capacity, there were no significant differences among the three media, although media A and B showed a tendency towards higher values (Figure 4B).

3.4. Gene Expression Analysis on the Calli Cultivated on the Three Best Media

Targeted gene expression profiling was carried out on calli grown on the three selected media (Figure 5). To determine whether medium-dependent differences existed in the expression of genes involved in phenolic compound biosynthesis, 12 genes encoding enzymes that participate in the phenylpropanoid pathway were screened. The genes studied are phenylalanine ammonia lyase (PAL), 4-coumarate:coenzyme A ligase (4CL), chalcone isomerase (CHI), flavone synthase (FNS), and flavonoid 3′-hydroxylase (F3′H3). Prenyltransferases (PT1-3-5-6) and O-methyltransferases (OMT6 and 21) were also included.

The genes PAL4, PAL7 and 4CL1 showed significantly higher expression in medium A than in medium C. The gene OMT6 showed significantly increased expression in medium B compared to medium C. On the other hand, the genes CHI2 and F3′H3 showed higher expression in medium C compared to medium A. No statistically significant differences could be observed for the prenyltransferases PT1-3-5-6 and the O-methyltransferase OMT21.

3.5. Cultivation of Other C. sativa Genotypes on the Three Selected Media

In order to verify whether the three selected media allowed for the successful propagation of calli from other genotypes, one non-commercial genotype (Local 14; [23]) and one commercial genotype (Fédora 17) were also investigated (Figure 6). After 1 and a half months of culture, medium A showed more vigorous and friable calli compared to media B and C. These differences were particularly evident in the commercial genotype Fédora 17, which showed extensive browning on medium C (Figure 6B).

4. Discussion

The interest in C. sativa tissue culture has increased substantially in recent years [15,18,38,39,40,41,42,43,44], driven by the potential to produce high-value bioactive compounds. The well-characterised phytochemical profile and economic importance of C. sativa make it an ideal model for optimising bioprocesses, elicitation strategies, and advanced propagation techniques, with implications for pharmaceuticals, breeding, and conservation strategies. In this respect, the present study confirmed that PGRs play a decisive role in callus induction and growth. We focused on two major PGR classes, auxins and cytokinins, both of which are well recognised as key regulators of callogenesis in C. sativa [13,27,44]. Their synergistic interaction promotes cell dedifferentiation and proliferation, enabling callus initiation and sustained growth [10].

Monthony et al. [27] provided a comprehensive overview of callogenesis protocols in C. sativa, highlighting the recurrent use of kinetin, NAA, 2,4-D, BAP, and thidiazuron (TDZ) in different combinations. These studies also emphasised the strong genotype dependence of callogenesis, underlining the need for systematic optimisation of PGRs for each genetic background.

As an initial step, this study evaluated 49 PGR combinations to identify an efficient medium for C. sativa callus induction and long-term maintenance, with potential applicability to the subsequent establishment of cell suspension cultures. The most effective callogenesis media were identified by assessing different parameters: percentage of callus induction, extent of callus coverage across the stem, callus proliferation, colour, and texture. Among the auxin-cytokinin combinations tested, the media supplemented with 2,4-D 4.5 µM + kinetin 1.5 µM (medium A), 2,4-D 1.5 µM + kinetin 1.5 µM (medium B), and 2,4-D 1.5 µM + BAP 1.5 µM (medium C) gave the best results (Figure 2). Subsequent assessment of total phenolic content, antioxidant activity and gene expression further distinguished these media, with 2,4-D 4.5 µM + kinetin 1.5 µM emerging as the most effective in terms of calli vigour, colour, texture, total phenolic content and expression of genes intervening in the phenylpropanoid pathway (Figure 4 and Figure 5).

In C. sativa, the majority of published studies have focused on regeneration protocols, frequently using TDZ combined with NAA or 2,4-D and often employing drug-type genotypes [45]. By contrast, our study specifically targeted callogenesis and callus proliferation in a non-drug and non-commercial genotype, which could explain the observed differences in response. Mechanistically, although both NAA and 2,4-D are synthetic auxins, they trigger distinct downstream pathways. As shown by Campanoni and Nick [46], 2,4-D efficiently activates the auxin receptor-mediated program that promotes cell division, whereas NAA preferentially stimulates cell elongation. This can explain why 2,4-D combinations generally produce a higher callus proliferation. Higher concentrations of NAA are required to achieve comparable proliferation [46].

In the set-up studied here, the combination of 2,4-D with kinetin outperformed NAA-based treatments (Figure 2 and Table 2). The consistent effectiveness of this combination across various species [47,48,49] supports its general utility; nevertheless, the genotype-dependent variability in growth observed in the two other C. sativa genotypes, the non-commercial Local 14 and the commercial Fédora 17 (Figure 6), highlights the importance of tailored optimisation in this species.

The capacity of these PGRs to influence callus texture/colour/vigour, as well as the abundance of total phenolic compounds, the antioxidant capacity and the expression of genes intervening in the phenylpropanoid pathway, suggests that careful medium selection can simultaneously optimise both morphological and metabolic outcomes, which are essential for the establishment of robust bioprocesses aimed at producing secondary metabolites. Several reports have described the effect of PGRs on phenolic accumulation in both in vitro-cultivated plants and calli [50,51,52,53].

Gene expression analysis revealed a significantly higher expression of genes involved in the upstream steps of the phenylpropanoid pathway, i.e., PAL4, PAL7, and 4CL1, in medium A containing 2,4-D 4.5 µM + kinetin 1.5 µM. The increased expression of PAL genes in media containing kinetin is in agreement with the results reported in kinetin-treated tobacco callus [54], in which the transcriptional activation was accompanied by an enzymatic activity increase. In this study, the increased accumulation of total phenolics observed in callus cultures on media A and B (Figure 4) correlated with higher expression of PAL transcripts.

PAL catalyses the first committed step of the phenylpropanoid pathway, directing the C flux from primary metabolism towards the synthesis of diverse phenolic compounds [55]. Since PAL determines the entry rate into this pathway, it is legitimate to speculate that high PAL expression under the supplementation of 2,4-D and kinetin leads to increased PAL activity and enhanced phenolic biosynthesis. The higher PAL expression detected under the 2,4-D + kinetin treatment likely increased substrate availability for downstream enzymes such as 4CL, OMT, and FNS, contributing to the observed higher content in phenolic compounds. Although not statistically significant, media A and B, both based on 2,4-D and kinetin, showed a tendency towards higher expression values of FNS6 and OMT21 compared to medium C (Figure 6).

In contrast, CHI2 and F3′H3 expression reached their highest levels in medium C containing 2,4-D 1.5 µM + BAP 1.5 µM, indicating that BAP could favour the final conversion steps leading to diverse flavonoids. Indeed, CHI leads to the biosynthesis of naringenin and F3′H to the biosynthesis of quercetin or luteolin [36,56].

The differences observed in gene expression when using different cytokinins are not surprising: transcriptomic comparisons on leaf tissues of Arabidopsis plants showed that kinetin and BAP activate distinct gene sets, including transcripts intervening in the phenylpropanoid pathway [57]. In A. thaliana leaves, BAP induced the early steps of phenylpropanoid synthesis and repressed almost all flavonoid/anthocyanin biosynthetic genes. The C. sativa callus model used here revealed a kinetin-associated upregulation of upstream biosynthetic genes, suggesting that cytokinin identity modulates phenolic metabolism in a context-dependent manner. Indeed, the experimental system employed in this study differs substantially from leaf tissues, and such differences in developmental context likely account for the divergent regulatory responses observed.

This study presented a comprehensive approach to optimising callogenesis in C. sativa through the screening of 49 auxin–cytokinin combinations and identified the most suitable one for callus maintenance and production of phenolic compounds. Beyond optimising callus induction, future research should focus on developing stable C. sativa cell suspension cultures to enable scalable biotechnological applications such as secondary metabolite production. Moreover, assessing the genetic stability of long-term cultures through molecular markers or epigenetic profiling will be essential to ensure the fidelity and reproducibility of in vitro-derived cell lines over time.