The establishment of micropropagation protocols for a particular species, cultivar or strain often precedes the utilisation of genetic engineering and gene modification technologies as they offer alternative routes to produce interesting metabolites that are of industrial value. With this in mind, several different strategies are available for genetic manipulation despite the plant showing recalcitrance to in vitro regeneration [94] and genetic transformation [95]. Agrobacterium-mediated transformation has been explored by Feeney and Punja [94] and the production of hairy root cultures of five different strains (namely, Futura77, Delta-llosa, Delta405, CAN0111 and CAN0221) were established using seedling leaf material, and hypocotyls, and cotyledons to produce both axenic callus and root cultures using the Ti and Ri plasmid systems, respectively [95]. Such cultures are thus amenable to use as scientific tools to study the biosynthesis pathways associated with Cannabis and for their application in industrialised production of Cannabis-derived metabolites. Experimentation with the Agrobacterium co-cultivation strategies to enable for successful production of transgenics is an important consideration [96]. Of the strains tested— A4, ATCC15834, MSU440, and A13 (MAFF-02-10266)—the A. rhizogenes MSU440 strain was the most efficient in generating various rol-transformed transgene lines. Successful genetic transformation using Agrobacterium-mediated technology is often challenged by different responses associated with using a variety of different strains and the study by Sorokin et al. [97] showed positive transient expression when seedling explants were transformed with a strain of A. tumefaciens (EHA105) harbouring the pCAMBIA130-uidA GUS reporter gene. Although many of these studies aim to ultimately increase metabolite yields in plant systems, other researchers and biotechnology companies intend to focus on using microbial systems for the heterologous manufacture of desired compounds for en masse production of cannabinoids that occur in nature in trace quantities [98]. With this in mind, it becomes critical to better understand the regulatory mechanisms that control specialised metabolism in Cannabis.

The application of CRISPR Cas9 gene editing techniques also holds promise for use in both C. sativa and C. indica and with future applications being imminent as CRISPR Cas 9-mediated editing has several advantages for Cannabis metabolite engineering. As detailed by Deguchi et al. [98], gene knockouts that target multiple branch pathways may alter metabolic flux, leading to increased production of cannabinoids and terpenoids of pharmaceutical interest. Base editing of single-nucleotide polymorphisms associated with the Cannabis genome also offers superior methods that will likely have a higher public acceptance than conventional genetic modification strategies for the production of elite chemotypes of C. sativa [98]. Such technologies are also highly suitable for molecular breeding for traits not only associated with phytochemical composition of Cannabis strains but also to impart better disease and pest resistance, polyploidy manipulation plus the alterations to general patterns of plant and growth development of Cannabis plants [83,98].

Broadening the scope of using gene manipulation techniques in Cannabis has met with some stumbling blocks despite major advancements in determining the regulatory and biosynthesis genes that control specialised metabolic pathways of key medicinal metabolites uniquely produced by Cannabis that could be used for pathway engineering [83]. The main obstacle hindering successful gene engineering is associated with the inability to produce stable transgenics that proceed past the transient transgene expression phase as available plantlet regeneration procedures are unfortunately not presently available for many genotypes of Cannabis. The most comprehensive study thus far in solving these problems is that of Zhang et al. [83], where the authors tested 100 hundred strains of hemp obtained from the national germplasm bank of Institute of Bast Fiber Crops (IBFC). This was mainly motivated by the fact that genotype-specific transformations for Cannabis are limited in their application across many commercial varieties.

CRISPR-Cas9 mutagenesis and Agrobacterium-derived genetic transformation with several genes (ZmWUS2, NbSTM, NbIPT, OsGRF4 and AtGIF1) that promote somatic embryogenesis and act as gene regulators of pathways that control organogenetic plant responses in other model plant species were explored therein. Although the hemp strain referred to as YUNMA7 is a commercial variant that is most important in China was included in the experiments, the best Cannabis strain was determined to be the DMG278 type. With the aim of genetic transforming hemp, this was then chosen as the main experimental model as it was more amenable to Agrobacterium-mediated genetic transformation using an AGL1 strain. Transgenics produced in this way were ultimately grown as F2-generation transgenic crops in the field. Together with the production of transgenics, gene editing experiments used a protoplast-based method that was tested initially using six different guides for the CsPDS1 target locus. Although the production of viable protoplasts was at times inefficient, a cocktail of enzymes, in particular, 12 g/L pectinase and 5 g/L cellulose R10 enzymes, allowed for a greater pool of protoplasts that were transformed with the pAT-GFP plasmid. The explant choice, transformation with developmental regulators and the use of gene editing tools as a combined set of processes enabled for the regeneration of hemp transgenics to be successful, setting a new paradigm for exploitation in functional genomics research and for using synthetic biology approaches for improving Cannabis plants [83]. In another study, a protoplast-transformation technique using in vitro plantlets of the strain Cherry x Otto II, that inherently produces low THC and high levels of CBD, was used as starter material to introduce the DR5:GFP auxin-sensitive reporter gene together with the p35:RFP expression cassette. Flow cytometry analyses showed GFP-linked fluorescence of the heterologous Arabidopsis auxin-responsive element, eliciting IAA-related changes in transgenic protoplasts [99].

According to Ahmed et al. [100], plant genetic engineering that employs nanoparticle technology presents novel and revolutionary approaches that can overcome challenges associated with plant species that have proven recalcitrance to Agrobacterium-based gene modifications. In addition to this, nanoparticle-assisted transformation may utilise passive infiltration of plant tissues, making it easier to introduce multigene constructs into target tissues that would otherwise require sequential transformation steps with Agrobacterium. The latter method may thus become labour intensive and be plagued by poor transformation frequencies when the purpose is to introduce multiple genes into target tissues. As a proof of concept, Ahmed et al. [100] showed transient gene expression in Cannabis of two soybean transcription GmMYB29A2 and GmNAC42, cloned into the pGWB6 plasmid, that control the synthesis of glyceollin, which functions as a defence metabolite. Lower leaf surfaces of the Tygra variety were passively infiltrated with poly-ethylenimine cationic polymer-modified silicon dioxide-coated gold (PEI-Au@SiO₂) nanoparticles using a multigene delivery system for the MYB transcription factors that were linked to a GFP expression marker. Attempts to use Agrobacterium-mediated transformation are ongoing as researchers try to expand the varieties that are being tested with this method. Concomitantly to assessing the effectiveness of agrobacterial co-cultivations with seedlings of the short-day C. sativa agricultural lines (i.e., Ferimon, Felina32, Fedora17, USO31 and Futura71) and the day-neutral FINOLA type with the intention of generating transgenic plants, an in vitro regeneration protocol was assessed when plants were placed on 0.2 mg/L NAA: 4 mg/L TDZ MS medium [101]. Although cotyledons, hypocotyls and apical meristem explants were co-incubated with Agrobacterium tumefaciens LBA4404 (pBIN19), as expected, not all these explants had similar transformation rates. Hypocotyls were most amenable to Agrobacterium-mediated gene modification, recorded at 53.3%. Some regenerants had an albino phenotype on the 100 mg/L kanamycin selection medium but spontaneous in vitro regeneration and rooting were evident in certain transformed lines. This study by Galán-Ávila et al. [101] also indicated that the Futura75 strain had the highest transformed hypocotyl-derived transgenics that were transiently expressing the GUS reporter gene, whereas for the FINOLA, Fedora17 and Felina52 varieties, no transgenics were produced, further emphasising on the genotype-dependent specificity aspect that is integral to biotechnological manipulations of Cannabis.

The difficulties in obtaining transgenic plants of this plant continue to hinder scientific progress in studying its genes using functional screening assays, making it impossible to gain deep insights into the regulatory mechanisms that control specialised metabolite production in Cannabis. Because the unique enzymes that are responsible for cannabinoid synthesis have been elucidated using genome sequencing tools, regulation of this pathway is highly interesting for biotechnologists and natural product researchers alike. Protoplast transformation appears to be a more dependable and viable pathway for greater number of strains. With the intention to use this method for CRISPR/Cas9 gene editing, to study protein–protein interactions and other genetic-biochemical regulatory mechanisms that control cannabinoid metabolism, Matchett-Oates et al. [102], in a recent study, targeted a high-THC-producing C. sativa genotype to examine PEG-mediated protoplast transformation from leaf mesophyll cells. The transformation was confirmed to be correlated to the plasmid and PEG concentrations to elicit transient GFP expression. Even though PEG-assisted protoplast transformation was optimised for a high THC-containing strain, two other strains (Cannbio2 and a high CBD strain) responded adequately with protoplasts that could also be used for future gene editing manipulations and other molecular-based scientific investigations.

It is thus clear that the possibilities that genetic engineering and gene editing present for furthering our general understanding of cannabinoid metabolism and its regulation provide strong incentives to continue researching simple, less labour intensive, viable protocols for producing a broad-spectrum of genetically modified Cannabis strains. Such protocols, that can be routinely applied across many different laboratories, are most likely imminent as this research area has gained considerable attention from biotechnologists with a vested interest in turning Cannabis into a model species to study its metabolism.

Non-transgenic methods using chemical mutagens are also suitable for polyploidy generation in Cannabis and a tissue culture step is often a prerequisite for the generation of plants with different chromosome composition than the mother stock [103]. As an example, Parsons et al. [65] generated Cannabis polyploids with higher CBD content and sesquiterpene accumulation via tissue culture albeit rooting was lowered in the tetraploids developed using oryzalin, a compound that alters microtubule formation during mitosis. Another example is illustrated by the work of Kurtz et al. [103] using colchicine as the chemical mutagen to induce increased number of chromosomes in in vitro-derived lines from germinated seedlings. A tetraploid hemp line was chosen as having the best potential for downstream breeding applications despite the likelihood of inherent genetic variation that may be produced in some seed.

Female plants are favoured in Cannabis farming, with feminised seed being highly sought after. Polyploids are useful for plant breeding as more desired phenological traits and quality assured phytochemical traits may be expressed in bred plant lines [103]—such manipulation is of high relevance for producing elite Cannabis variants. 5. Conclusions and Future Prospects

Micropropagation via direct organogenesis or indirect organogenesis may be a useful tool in propagation of Cannabis sativa for mass production of the crop yields with high vigour whilst space requirements are minimised for mass amounts of plant clones. Direct organogenesis allows for genetic preservation of the plants, producing genetically identical clones that are vital for industry use. Micropropagation of Cannabis is a fairly new field that is, however, rapidly developing with the opening of new industrial markets globally that have been spurred on by the legalisation of Cannabis derived products. Although there have been protocols developed, many more trials and combinations of PGRs experiments are needed to gain further insight into the development of an appropriate species-specific micropropagation regimen that elicits highly prolific organogenesis in culture of morphologically normal and healthy plants. The exploration of different strains of Cannabis is also still at its infancy and available protocols are not necessarily highly efficient as they are plagued by low regeneration capacity. To reiterate, more strain-specific investigations are thus also urgently needed to provide a better understanding of strain-related differences with respect to in vitro culture. There is thus room to investigate methods that will allow for efficient, reliable and easy-to-reproduce regimes that are characterised by high multiplication rates and rapid ex situ plant establishment under greenhouse and/or field conditions. Many of the micropropagation regimes discussed in this review do not necessarily include a post-culture analysis of the phytochemistry once plants have been acclimated. This is important for quality control purposes if the application of in vitro cultivation techniques is deemed to maintain the chemical integrity of desired commercial chemotypes. Several different approaches may be useful in this regard including microsatellite analyses to determine genetic-associated somaclonal variation resulting from microculture conditions. Epigenetic effects that illicit unprecedented genetic changes have the potential to alter chemical fingerprints of cultivated Cannabis and, together with high-throughput metabolite fingerprinting, this may be useful in monitoring such changes as these can become heritable in subsequent generations. This is particularly important as most Cannabis is used for an industry dependent on phytochemical consistency. In the future, it is likely that cryopreservation as a germplasm conservation strategy may be applied for long-term preservation of unique chemotype variants of C. sativa. The preservation of heirloom seeds using synthetic seed technologies may be complimentary in the augmentation of breeding activities that solely focus on Cannabis and its relatives.

In addition, when the issues associated with difficulties in propagating Cannabis in vitro are no longer an impediment to the generation of healthy and true-to-type clonal propagules, many other genetic engineering applications including the use of fast-evolving gene editing tools will come into routine application in many different laboratories and biotechnology start-ups. This will lead to a revolution in the innovations associated with both medicinal marijuana and hemp.