Garcinia kola: a critical review on chemistry and pharmacology of an important West African medicinal plant

Garcinia kola Heckel (Clusiaceae) is a tree indigenous to West and Central Africa. All plant parts, but especially the seeds, are of value in local folklore medicine. Garcinia kola is used in treatment of numerous diseases, including gastric disorders, bronchial diseases, fever, malaria and is used to induce a stimulating and aphrodisiac effect. The plant is now attracting considerable interest as a possible source of pharmaceutically important drugs. Several different classes of compounds such as biflavonoids, benzophenones, benzofurans, benzopyran, vitamin E derivatives, xanthones, and phytosterols, have been isolated from G. kola, of which many appears to be found only in this species, such as garcinianin (found in seeds and roots), kolanone (fruit pulp, seeds, roots), gakolanone (stem bark), garcinoic acid, garcinal (both in seeds), garcifuran A and B, and garcipyran (all in roots). They showed a wide range of pharmacological activities (e.g. analgesic, anticancer, antidiabetic, anti-inflammatory, antimalarial, antimicrobial, hepatoprotective and neuroprotective effects), though this has only been confirmed in animal models. Kolaviron is the most studied compound and is perceived by many studies as the active principle of G. kola. However, its research is associated with significant flaws (e.g. too high doses tested, inappropriate positive control). Garcinol has been tested under better conditions and is perhaps showing more promising results and should attract deeper research interest (especially in the area of anticancer, antimicrobial, and neuroprotective activity). Human clinical trials and mechanism-of-action studies must be carried out to verify whether any of the compounds present in G. kola may be used as a lead in the drug development.


Introduction
Garcinia kola Heckel (Clusiaceae), a multipurpose tree commonly found in subtropical and tropical moist lowland forests of Nigeria, Cameroon and other countries in sub-Saharan Africa. It is colloquially called bitter kola, false kola or sometimes ''wonder plant'' because almost every part of this tree has been used in traditional medicine for broad portfolio of ailments since ancient times (Ijomone and Obi 2013;Maňourová et al. 2019;Erukainure et al. 2021). The seeds are highly valued as oral masticatory agent with bitter astringent taste and stimulant effect. They (as well as other plant parts) are used to treat wide range of diseases, including as gastric and liver disorders, diarrhoea, bronchial diseases, throat infections, colds, fever, malaria, and as an aphrodisiac (see Table 1) . Especially the use in the area of liver protection and disease, throat infection, colds, and the aphrodisiac action is often repeated in the literature. The seeds are habitually chewed as a part of traditional, cultural and social ceremonies and for their aphrodisiacal effect (Maňourová et al. 2019). It is often given to guests and unfamiliar persons as sign of friendship and respect. Currently, G. kola is recorded as ''vulnerable'' in IUCN's Red List of Threatened Species, possibly due to deforestation practices and relatively intensive collection from the wild (Cheek 2004).
Over the last few years, G. kola has received quite large research attention, mainly due to content of a very specific biflavonoid complex collectively referred to as kolaviron, whose distribution seems to be limited to G. kola. More recently, this research attention has resulted in emergence of a few review articles (Maňourová et al. 2019;Erukainure et al. 2021;Dogara et al. 2022;Emmanuel et al. 2022) that introduce G. kola, and kolaviron, as a promising material for drug discovery. However, kolaviron is not the only constituent found in G. kola and it contains other compounds (e.g. garcinianin, kolanone, gakolanone, garcinoic acid, garcinal, garcifuran A and B, and garcipyran) (Hussain et al. 1982;Niwa et al. 1993Niwa et al. , 1994aTerashima et al. 1995Terashima et al. , 1997Akoro et al. 2020) that also appears to be very specific for G. kola and their presence have thus fur not been confirmed in any other botanical source. These compounds are to a very large extent neglected in these review articles and kolaviron is perceived as the active principle, though these lesser-known compounds may provide interesting pharmacological activities as well. On top of that, kolaviron is in majority of available studies (animal models) tested in very large doses, which appears rather unrealistically high and untransferable to clinical practice. Clinical data on humans on any of the constituent found in G. kola are entirely missing. Despite of this fact, these reviews draw conclusions on therapeutic efficacy of G. kola and kolaviron. In view of what is written above, this review offers a critical update on available information of the most studied and discussed compound of G. kola, kolaviron, and provides analysis of existing knowledge on other present constituents.

Chemical composition
Primary metabolites Although G. kola seeds are more valued for their medicinal properties rather than as foodstuff, the kernels are still commonly consumed, which justifies concerns about their nutritional value (Okoye et al. 2014). There are wide discrepancies among the published results on the species primary metabolites content. Generally, the studies agree on relatively high amounts of moisture in the seeds (about 70%), suggesting their vulnerability to mould infestation and possible storage/post-harvest processing difficulties. Present saccharides, also described as nitrogenfree extracts (NFE), form the largest part of the seed proximate composition (around 65%), while the content of minerals is very low (1.5% on average). The mean value for crude protein was found to be 3.5%, with lysine (2.4 g/kg), leucine (1.9 g/kg) and valine (1.7 g/kg) being the predominant essential amino acids (AA) and glutamic acid (6.8 g/kg) with arginine (5.5 g/kg) as the highest abundant nonessential AA in both kernels and seeds' hulls (Eleyinmi et al. 2006). The crude fat generally varies about 6.2% with oleic acid (C 18:1; 38 mg/kg), linoleic acid (C 18:2; 36 mg/kg) and palmitic acid (C 16:0; 32 mg/kg) being the dominant fatty acids in both seeds and hulls (Eleyinmi et al. 2006). The crude fibre content was determined at 9.4% on average. Before consumption, people generally prefer to peel the seeds, discarding the hulls as worthless waste. However, due to their high protein content (9.92 g/100 g), these husks may represent a valuable fodder source for domestic animals, whose diet is usually based only on natural pastures of poor quality and thus quite low in protein content (Eleyinmi et al. 2006). If grinded into a powder, the hulls can be incorporated into enriched feeding mixtures.
Quite limited information is available on the micronutrient content of G. kola seeds. They were reported to contain relatively high amounts of vitamin C (23.1-69 mg/100 g), potassium (25-722 mg/kg) and phosphorus (3.3-720 mg/kg) (Okwu 2005;Onyekwelu et al. 2015). They are also low in antinutrients such as phytate and oxalate, and are thus considered safe for consumption without any reports on harmful overdosing (Onyekwelu et al. 2015;Konziase 2015).

Secondary metabolites
Various classes of secondary metabolites have been isolated from different plant parts of G. kola. Of these, perhaps the most studied are flavonoids and their related structures. Benzophenone, benzofurans and benzopyran analogues, vitamin E derivatives, xanthones and phytosterols have also been isolated from G. kola in the past. Many of the present constituents, namely, kolaviron, garcinianin, kolanone, gakolanone, garcionic acid, garcinal, garcifuran A and B, and garcipyran A, appear to be exclusive for G. kola and have not been thus found in any other plant species yet. A list of known compounds isolated from G. kola, including plant parts where these constituents have been found, are given in Table 2. Their corresponding structures are illustrated in Fig. 1.
At least seven biflavonoid structures have been characterized in G. kola, the most known and studied being kolaviron. Kolaviron is the principal biflavonoid mixture in the seeds and constitutes of biflavonoids GB1 (1), GB2 (2), and kolaflavone (3) (Ijomone and Obi 2013). Some authors confirmed the presence of kolavirone in the roots as well (Iwu et al. 1990c). Seeds and roots were also found to contain garcinianin (4) (Terashima et al. 1995;Ajayi et al. 2014). It appears that both kolaviron and garcinianin are exclusively produced by G. kola and are not found in other Garcinia species. Garcinia kola seed also contains amentoflavone (5) (Iwu and Igboko 1982); which is quite abundantly distributed across plant species (e.g. Gingko biloba and Hypericum perforatum) (Lobstein-Guth et al. 1988;Baureithel et al. 1997). Other biflavonoids occurring in G. kola include volkensiflavone (6) and morelloflavone (fukugetin; 7); so far they were only identified in the wood (Acuña et al. 2012). On the other hand, both compounds were also discovered in fruits of other Garcinia species (e.g. G. spicata, G. xanthochymus, G. intermedia, G. livingstonei, G. hombroniana), suggesting that they also occur in the fruits and seeds of G. kola. The benzophenones in G. kola are represented by kolanone (8), gakolanone (9), and garcinol (10). Kolanone was the first discovered benzophenone derivative in G. kola. It was found in various plant parts, including the fruit (Hussain et al. 1982), seeds (Madubunyi 1995;Uwagie-Ero et al. 2020), and roots (Iwu et al. 1990c).
As with biflavonoid kolavirone, the distribution of kolanone appears to be limited to G. kola and its presence has not yet been demonstrated in any other species. One recent study also confirms occurrence of structurally related gakolanone in G. kola stem bark (Akoro et al. 2020). It as well seems to be restricted only to G. kola. It was also discovered that the roots  Terashima et al. (1995Terashima et al. ( , 1997, Ajayi et al. (2014) 5 Amentoflavone Seeds, wood Iwu and Igboko (1982) 6 Volkensiflavone Wood Acuña et al. (2012) 7

Biological activities of kolaviron (KV)
A brief description of the biological activities of KV is given below; detail description (disease, dose, mode of administration, etc.) is given in Table 4.

Hepato-, nephro-, and gastrointestinal-protective activity
Hepatoprotective effect is one of the major area where KV was tested. The biflavonoid was investigated in animal models to protect the liver from a broad spectrum of hepatotoxic agents. Despite the intensive research, the exact mode of hepatoprotective action of

Xanthones
KV is still not fully understood. Some authors proposed direct antioxidant mechanism (e.g. via KV's ability to scavenge free radicals) (Alabi and Akomolafe 2020), while others pointed out that KV enhances activity of drug-detoxifying enzymes (KV increases the activity of UDP-glucuronosyl transferase and glutathione S-transferase) (Olatunde . Farombi et al. (2009) also suggested that its effect may be achieved through inhibition of cyclooxygenase (COX) and inducible nitric oxide synthase (iNOS) expression. Similarly, KV was also tested in the scenario of renal (Adaramoye 2009;Adedara et al. 2015;Offor et al. 2017;Alabi et al. 2018) and gastro-intestinal (Olaleye and Farombi 2006;Onasanwo et al. 2011;Akinrinde et al. 2015) protection in animal models against similar toxic agents as in the case of liver toxicity tests. Both nephroprotective and gastro-protective effect is presumably exerted via similar mode of action. Apart from mechanisms discussed above, it was also suggested that KV interferes with regulation of such structures as C-reactive proteins (CRP) and extracellular signal regulated kinase (ERK) (Ayepola et al. 2014b;Akinrinde et al. 2016;Oyagbemi et al. 2018b).
In the case of gastrointestinal protective activity, KV was suggested to inhibit proton pump, thus providing anti-ulcerogenic effect.

Effect on heart and cardiovascular disorders
In early studies, KV was shown to produce hypolipidaemic effect and to reduce the relative heart weight of cholesterol-fed rats. Its activity was comparable to that of cholestyramine (questran), a commonly used hypocholesterolemic drug (Adaramoye et al. 2005). Additionally, KV was found to lower blood pressure in hypertensive rats (Uche and Osakpolo 2018;Olatoye et al. 2021). In other studies dealing with animal ischemic/reperfusion model, KV demonstrated to attenuate the heart injury through interference with apoptotic pathway (e.g. caspase reduction/cleavage), and reperfusion injury signaling kinase (RISK) (Oyagbemi et al. 2017(Oyagbemi et al. , 2018a. In a more recent study, KV also reduced cardiovascular injury in fructose-streptozotocin induced diabetic rats (Adoga et al. 2021). Furthermore, KV showed cardioprotective effect in animal models against various cardiotoxic agents, including antitumour drugs, and antimalarial agents (e.g. amodiaquine and artesunate) (Ajani et al. 2008).

Effects on central nervous system (CNS)
The early studies of KV were focused on in vitro determination of its protective activity against atrazine in certain neurological cell cultures (e.g. human dopaminergic SH-SY5Y and PC12 cells) (Abarikwu et al. 2011a, b). The CNS experiments were afterwards transferred to animal models, where KV showed neuroprotective effect against several neurotoxins. It was suggested that antioxidant effect (i.e. enhancement of antioxidant defences) might be the major mechanism of its beneficial action, though other modes were proposed as well (such as inhibition of stressor molecules and toxic proteins production). KV also demonstrated positive results in the animal models of cuprizone-induced multiple sclerosis. Again, its beneficial effect was explained by  intraperitoneally, i.d.
intradermally, N/A information not available antioxidant-related action (Omotoso et al. 2018a(Omotoso et al. , b, 2019. A neuroprotective effect was also observed in various rat models of CNS disorders. It was suggested, that KV might exert its neuroprotective effect through anti-inflammatory and antiapoptotic mechanisms. Additionally, KV was also suggested to be a potential inhibitor of acetylcholinesterase (AChE) (Ijomone and Obi 2013;Akinmoladun et al. 2018), though this was deduced only on the basis of reduced staining activity of AChE and not by enzyme-binding study. Moreover, very recently KV indicated an antiamyloid activity via destabilization of the assembled Ab particles in a molecular docking study (Adewole et al. 2021a).

Effect on reproduction and infertility
Garcinia kola is relatively widely used in traditional medicine as an aphrodisiac. Corresponding with this fact, studies have been focused on examining the effect of present substances on reproductive properties. KV was found to prevent testicular damage and decline of sex hormones upon administration of various toxic agents. Administration of these agents resulted in increased levels of antioxidant/detoxifying enzymes (catalase, superoxide dismutase, glutathione S-transferase) and markers of oxidation (e.g. elevated hydrogen peroxide and malondialdehyde). Additionally, the rats that had been treated with KV also showed improved semen characteristics (e.g. sperm count). It was also found out, that KV has lowered the negative effect of EGEE on activities of 3b-hydroxysteroid dehydrogenase (3b-HSD) and 17b-hydroxysteroid dehydrogenase (17b-HSD), enzymes that are associated with production of steroidal hormones (e.g. testosterone) (Adedara and Farombi 2013).

Diabetes
Investigations were made to figure out if KV can act as a potential source of diabetes treatment. The compound showed a hypoglycaemic effect in alloxaninduced diabetic rabbits and streptozotocin-induced diabetic rats (Iwu et al. 1990b; Adaramoye and Adeyemi 2006; Adaramoye 2012). Though there is no generally accepted mechanism of action yet, KV was suggested to produce its antidiabetic effect via inhibition of a-glucosidase and a-amylase activities (Iwu et al. 1990b;Salau et al. 2020). Recent study has also suggested that KV may play a regenerative role in pancreatic islets in streptozotocin-induced diabetic rats (Oyenihi et al. 2021). Other studies focused on the KV ability to reduce secondary pathology complications associated with diabetes, including hepatoxicity, nephrotoxicity, and cardiotoxicity. These activities have been discussed in previous sections.

Pain and inflammation
Anti-inflammatory activity of KV was firstly studied in an carrageenan-induced paw oedema mice model (Olaleye et al. 2010;Tchimene et al. 2015a). According to the subsequent tests on cell lines, it was suggested that KV might be the most active antiinflammatory principle of G. kola, interfering with the normal production of pro-inflammatory mediators such as prostaglandins (via COX enzymes inhibition), nitric oxide, interleukins, tumour necrosis factor a (TNF-a), monocyte chemotactic protein-1 (MCP-1), and vascular endothelial growth factor (VEGF) (Olaleye et al. 2010 Additionally, KV was found to reduce neuroinflammation in BV2 microglia/HT22 hippocampal neuron co-culture, exerting its activity via the same mechanisms mentioned above. Also, KV showed to possess a certain analgesic effect (Tchimene et al. 2015b). It was later discovered that its pain-relieving activity may not probably be associated with COX-2 inhibition, but rather involves nitrergic and ATP-K ? sensitive pathways (Ibironke and Fasanmade 2015).

Immunomodulatory activity
There are a few studies addressing the immunomodulatory activity of KV. In the first report, research on immunocompetent and immunocompromised models in rats was carried out, where KV showed inhibition of delayed-type hypersensitivity and increase in the primary and secondary sheep erythrocytes-specific antibody titers. The results showed that administration of KV ameliorated the cyclophosphamide-induced leukopenia and increased the number of white blood cells (Nworu et al. 2008). In other studies, KV delayed the development of the clinical symptoms of influenza in the infected mice (Awogbindin et al. 2015). Apart from other mechanisms discussed above (e.g. inhibition of COX-2, interleukins, and cytokines production), it was suggested that KV is capable of fostering the CD4 ? response (Awogbindin et al. 2017).

Cancer
Only a few studies regarding KV anticancer activity exist, though vast majority of them aim specifically at determining effect on biochemical parameters of benign prostatic hyperplasia in rats. KV showed a similar effect on serum levels of prostate specific antigen, total prostatic proteins, prolactin, oestradiol, testosterone, testosterone/oestradiol ratio, urea, and creatinine as the control finasteride (Kalu et al. 2016;Winner et al. 2016). Since antiandrogen finasteride is an 5a-reductase inhibitor, it was suggested that KV has the same mechanism of action. Yet, the therapeutic efficacy of KV in benign prostate hyperplasia (BPH) are far from conclusive. On top of that, it is worth to note that if indeed KV was an 5a-reductase inhibitor it would contradict the traditional aphrodisiac ethnomedicinal indication of G. kola. Recently, KV was also found to protect U937 cell and macrophages from bromate-induced cytotoxicity in an in vitro study (Okoko and Ndoni 2021). Moreover, histone deacetylase inhibitory activity was shown in an in silico model (Adewole et al. 2021b).

Antiparasitic activity
Although G. kola is commonly used in folk medicine to treat malaria, there are relatively few studies on its antimalarial effect. KV showed anti-malarial activities by suppressing Plasmodium bergheii in infected laboratory mice (Oluwatosin et al. 2014;Tshibangu et al. 2016). Of all KV components, GB1 exhibited the almost the same in vitro antimalarial effectivity on P. falciparum as quinine. In the in vivo test, it was observed that GB1 significantly increased the average life span of Plasmodium-infected mice (Konziase 2015). Recently, KV was also showed to be effective against Trypanosoma infections (e.g. T. congolense) both in vitro and in vivo. It has been suggested that KV may exert its antitrypanosomal activity by interfering with trypanothione reductase, an enzyme responsible for homeostasis maintenance (Timothy et al. 2021).

Anti-snake venom activity
Anti-snake venom activity forms a relatively narrow area of KV research. As far as we know, only one study addressed this issue. Quite recently Okafor and Onyike (2020) suggested that the KV may produce inhibitory effect against hydrolytic enzymes of Naja nigricollis venom, namely phospholipase A2 (PLA2), protease, hyaluronidase and l-amino acid oxidase, and thus also neutralize their myotoxic, oedemic, haemolytic and procoagulant effects. However, KV was assayed at quite high doses (venom:KV 1:5 w/w) and reasonable inhibition was only observed in the case of PLA2. It is questionable whether these high doses of KV are clinically relevant.

Biological activities of amentoflavone
Amentoflavone is a widely studied biflavonoid. It is quite abundant in nature across various plant families, including-Ginkgoaceae, Selaginellaceae, Cupressaceae, Euphorbiaceae, Podocarpaceae, and Calophyllaceae (Yu et al. 2017). The main area where amentoflavone has been studied is anti-inflammatory, antitumour, antidiabetic, antifungal, antiviral, and neuro-and cardio-protective activities. Amentoflavone was found to interfere with levels of inflammatory mediators (e.g. nitric oxide, malondialdehyde, reduced glutathione, tumour necrosis factor alpha (TNF-a), and prostaglandin E-2) in various lipopolysaccharide-stimulated cell lines (Ishola et al. 2013). Additionally, amentoflavone was also reported to inhibit the production of proinflammatory interleukins, including IL-1b and IL-6 (Abdallah et al. 2015). As of yet, precise mode of its anti-inflammatory action has not been established. Amentoflavone have been tested for cytotoxic effect against various cancer cell lines. Several mechanisms of its anticancer action have been proposed, including induction of cell cycle arrest, apoptosis (e.g. interference with caspase-3), inhibition of fatty acid synthase (FASN) and phosphorylation of protein kinase B (PKB), and downregulation of HER2 protein (Lee et al. 2013). Amentoflavone was also suggested to regulate glucose level, production of insulin and to possess pancreasregenerating properties in diabetic mice (Su et al. 2019). It was indicated that it may exert its antidiabetic effect by inhibiting protein tyrosine phosphatase 1 (PTP1B) (Na et al. 2007). Amentoflavone demonstrated neuroprotective effect in various experiments. This activity may be related to interference with the receptors for serotonin, adrenaline, and GABA. Amentoflavone also showed protective activity against cardiovascular dysfunction in high fructose and fat diet induced metabolic syndrome rats. Administration decreased systolic blood pressure, left ventricular internal diameter and posterior wall thickness in diastole, increased fractional shortening and decreased ejection fraction, relative wall thickness, estimated left ventricular mass, cardiac stiffness and wet weight (Qin et al. 2018). Amentoflavone was also shown to reduce lipid accumulation and oxidized low density lipoprotein (ox-LDL) uptake in HUASMCs and THP-1 cells. It was suggested that amentoflavone acts as an inhibitor of proliferator-activated receptor gamma (PPARc) protein/cluster of differentiation 36 (CD36) signaling pathway (Zhuang et al. 2021). Additionally, amentoflavone was found out to inhibit phosphodiesterase in rat adipose tissue (Saponara and Bosisio 1998). Amentoflavone also showed antimicrobial activity against various fungal pathogens, including Candida albicans, Saccharomyces cerevisiae, and Trichosporon beigelii (Hyun et al. 2006 (Lin et al. 1997), and SARS-CoV 3CL pro (Ryu et al. 2010). As with the other biflavonoids mentioned in this review, amentoflavone, although possibly showing promising results in many of the in vitro and in vivo tests, has not yet been subjected to clinical trials and therefore its therapeutic efficacy is far from conclusive.

Biological activities of volkensiflavone/morelloflavone
It seems the biflavonoids volkensiflavone and morelloflavone display similar pharmacological properties as their related structure KV. However, the extent of research on them is far more limited. Their biological activities are summarized in Table 5. Perhaps the most widely studied area of these biflavonoids is antibacterial activity, though in the available in vitro studies, they are showing rather low efficiency. Both showed an ability to lower the minimal inhibitory concentration of norfloxacin against Staphylococcus aureus (E Silva et al. 2021). Anti-bacterial activity (e.g. against S. aureus, but also Bacillus subtilis, Pseudomonas aeruginosa and Escherichia coli) of volkensiflavone and morelloflavone and their glycosylated versions was also reported elsewhere (Trisuwan et al. 2013;Jamila et al. 2014). Interestingly, some of the glycosylated analogues of volkensiflavone and morelloflavone (e.g. (2R,3S)-volkensiflavone-7-O-bacetylglucopyranoside and (2S,3S)-morelloflavone-7-O-b-acetylglucopyranoside) did not demonstrate any notable antibacterial activity (Mountessou et al. 2018). Volkensiflavone demonstrated antiplasmodial activity in comparison to cholorquine. Morelloflavone showed approx. three-to ten-fold weaker activity than volkensiflavone (Azebaze et al. 2015). Contrastingly, Bezerra et al. (2021) reported morelloflavone to have activity against Leishmania amazonensis. Contrasting results were also found in another study, where volkensiflavone showed superior activity over morelloflavone against Leishmania infantum, but also other parasites (i.e. Trypanosoma brucei brucei and T. cruzi), while morelloflavone was not active at all (Mbwambo et al. 2006). Volkensiflavone and morelloflavone were also found to produce a vasodilatation via relaxation on aorta rings in a rat model. Both compounds were also suggested to be of interest as potential treatment for increased blood pressure and erectile disfunction (Brusotti et al. 2016). Moreover, their in vitro and in vivo atheroprotective effect caused by regulated interaction between oxidized low density lipoprotein (LDL) molecule and macrophages has been described as well (Tabares-Guevara et al. 2017).

Cancer
Garcinol is attracting a scientific interest mainly due to its ability to inhibit histone acetyltransferase (HAT), a novel drug target in cancer research. As a HAT inhibitor, garcinol was found effective at hindering the process of non-homologous end joining in the DNA repair mechanism, ultimately causing apoptosis of the cancer cells ( Table 6.

Anti-inflammatory activity
Garcinol disposed an anti-inflammatory activity in various animal models of induced inflammation. Majority of studies agree on a mechanism that appears to be related to the interference with NF-jB, iNOS, ERK, COX-2, p38 mitogen-activated protein kinases (MAPK), lipoxygenase (5-LOX), TNF-a, interleukin (e.g. IL-2, IL-6, IL-23), nuclear factor of activated T-cells (NF-AT) Schobert and Biersack 2019). Some authors also suggested that anti-inflammatory effect of garcinol is associated with HAT suppression (Ferriero et al. 2018).

Neurodegenerative disorders and drug withdrawal
Garcinol was found to be an inhibitor of monoamine oxidase B (MAO-B), and as such, it might be helpful in Parkinson's disease treatment by retarding dopamine depletion (Mazumder et al. 2018). Additionally, it was discovered that garcinol attenuated the sideeffects and increased bioavailability of L-DOPA, a dopamine precursor commonly used in the treatment of Parkinson's disease symptoms (Mazumder et al. 2016;Ryu et al. 2018). Garcinol also decreased mortality and seizure scores in mice, presumably by suppressing brain-derived neurotrophic factor (BDNF) and by having effect on neurotransmitter systems, including those involving glutamate and GABA A (Hao et al. 2016). Garcinol was also observed to decrease inflammation of microglia in rats via down regulation of NF-jB pathway and inhibiting COX-2, iNOS, and IL expression (Wang et al. 2017). A relatively unusual effect of garcinol has been discovered-in rats exposed to cocaine, garcinol inhibited restoration via reconsolidation-based modes following cocaine reactivation. The effect of garcinol on reactivated memories were long-lasting, suggesting a potential in control of drug abstinence and addiction (Fuchs and McLaughlin 2017).

Antiviral and antimicrobial activity
One of the early studies involved investigation on antiviral activity of garcinol against HIV, where again it was found to be potentially exerting its effect via inhibition of histone acetyltransferase (HAT) of the HIV infected cells (Mantelingu et al. 2007). Similarly as in the case of kolaviron, garcinol showed some degree of activity also against influenza virus (Hatakeyama et al. 2014). It appears that garcinol exerts its antiviral activity against influenza through regulation of the viral polymerase function (Schobert and Biersack 2019). Garcinol has demonstrated antibacterial, anti-yeast and antiprotozoal activity which was in some cases equal or better than conventional treatment. Again, mechanism of its antimicrobial effect might be related to the HAT inhibitory activity (noted above).

Biological activities of garcinoic acid (GA)
Compared to KV and garcinol, there is only a limited number of studies on GA. Its biological activities are summarized in Table 7. In the early reports, GA showed in vitro antioxidant (Terashima et al. 2002;Okoko 2009) and anticancer effect (Mazzini et al. 2009; Birringer et al. 2010). GA was also found to improve heart function in myocardial infarction rats by increasing levels of pro-angiogenic factors, including hypoxia-inducible factor 1-alpha (HIF-1a), VEGF-A, and basic fibroblast growth factor (bFGF) (Hu et al. 2020). Very recently, garcinoic acid, together with related structures garcinal and tocotrienol, have showed phosphodiesterase-5 (PDE-5) inhibitory activity in molecular docking study (Ojo et al. 2021). Their activity was comparable to sildenafil (ViagraÒ), suggesting that these compounds may be useful in treatment of erectile disfunction.
Additionally, GA was suggested to be of value as an agent with anti-inflammatory activity (Kluge et al. 2016). However, the exact mechanism of its antiinflammatory action has not yet been entirely established, though it was indicated that GA may act as a COX-2 and iNOS inhibitor (Wallert et al. 2019). GA also reduced deposition of Ab particles in brain in mouse a model of Alzheimer's disease (Marinelli et al. 2020). Once again, human clinical studies of GA are not yet available. As it is also noted above G. kola also contains quite large number of xanthone derivatives, which are also widely distributed throughout the higher plants. Other xanthone-producing species from the Clusiaceae family include Hypericum, Calophyllum, Kielmeyera and Tovomita (Terashima et al. 1999). These compounds are attracting a considerable research interest as they (e.g. various hydroxy-and methoxy-analogues) were reported to possess a wide range of biological activities, including anticancer, antimalarial, antimicrobial, anti-HIV, anticonvulsant, anticholinesterase, antioxidant, anti-inflammatory effect (Miladiyah et al. 2018;Ramakrishnan et al. 2021). It was also found out that xanthones may interfere with several enzymes, including a-glucosidase, acyl-CoA:cholesterol acyltransferase, aromatase intestinal P-glycoprotein, miRNA, protein kinase C, topoisomerase, and xanthine oxidase. Especially the hydroxyxanthones are thought to provide promising anticancer activity and their semi-synthetic variants are widely researched as anticancer drugs (Miladiyah et al. 2018). It was suggested that their anticancer effect may be connected to various mechanisms, including inhibition of cyclooxygenase-2 (COX-2), NF-jB pathway, cyclin-dependent kinases (Cdk), and by suppressing expression of vascular endothelial growth factor (VEGF) and metalloproteinases (Klein-Júnior et al. 2020). Since biological activities and pharmacology of xanthones are quite extensively given elsewhere (see references provided above) and they are not specific constituents of bitter kola and other Garcinia species, they are covered in this review only superficially. It must be noted though, that to the best of our knowledge, none of the xanthones have advanced into clinical use and are not medicinally used in treatment of any human disease. In addition to that, it appears that pharmacological properties in humans, including solubility, lipophilicity, dissociation constant, chemical and metabolic stability, permeability, transporters modulation, and plasma protein bindings are largely unknown for xanthones (Gomes et al. 2016). Even toxic effects are to a large extent unknown. Xanthones should be treated with caution as some of the very closely related compounds (e.g. aflatoxins) are known to produce pronounced toxicity to humans (Dewick 2009).

Toxicology
No serious adverse effects have been indicated in any of the available studies regarding the pharmacological activities of bitter kola. Few toxicological studies have been carried out which estimated the LD 50 of the bitter kola seeds to be as low as 5000 mg/kg b.w. (Okoye et al. 2014). Currently, the only perceived health risk of bitter kola is that consuming too much of the seeds can lead to fertility issues (Dogara et al. 2022). This may be related to the ability of the present compounds to alter sexual hormone levels through interaction with various enzymes (e.g. 5a-reductase). However, this activity have not been proved in a satisfactory manner (Kalu et al. 2016;Winner et al. 2016). Interestingly, it is in contradiction to the traditional aphrodisiac ethnomedicinal indication of G. kola. More toxicological studies and clinical trials of bitter kola and its compounds are required to elucidate this issue and to avoid any complications in connection with their possible clinical use.

Discussion
Majority of the available studies and review articles perceive KV as the active principle of G. kola. It has been investigated in a wide range of animal models and in vitro biological activities scenarios (see section biological activities of KV), where it was concluded that KV displays promising results in nearly every area studied and that it behaves almost like a panacea. A significant number of research articles are based on observing protective properties of KV against some toxin-induced disease model (e.g. kidney, liver, brain, heart, reproductive organs). Apart from very few exceptions, available studies have been using doses of KV ([ 100 mg/kg, 200 mg/kg and in some cases even [ 400 mg/kg), which may generally be viewed as excessive and from the clinical perspective unrealistic (calculated on a human body weight of 70 kg, the dose would correspond to administrations of approx. 7-14 g or even higher dose of pure substance). Even in the in vitro tests, it appears that in many cases way to high doses have been used ([ 10 lg/mL). Problematics of using too high doses in animal and in vitro studies is extensively reviewed in Gertsch (2009). When using such large doses in animal models, another question arises, and that is whether apart from beneficial effects one would also expect to observe adverse effects. No studies so far focused on possible side effects induced by overdosing of KV. However, from the available data from structurally related polyphenolics (e.g. resveratrol), it appears that high doses (e.g. approx. 2.5 g of pure substance consecutively for few days) are associated with nausea, vomiting, diarrhoea and liver dysfunction (Salehi et al. 2018). Additionally, in many studies, only one dose of KV has been tested, which does not allow an insight into how the substance behaves in a dosedependent manner and it does not provide statistical significance of a particular dose. Another problem with efficiency of KV lies in that very few studies have used appropriate control (many did not use positive control at all), and when they did so, it was used in an incomparable manner to the KV dose (often approx. 100-fold lower than that of KV). On the other hand, KV was in majority of studies administrated orally which corresponds with the traditional ethnomedicinal application.
Amentoflavone is another biflavonoid present in G. kola, whose research is even more extensive than that of KV (having more than 1000 references in scientific databases), yet it is very seldomly being associated with pharmacological properties of the species. Similarly, as in the case of KV, amentoflavone has been tested in many areas of pharmacological activity. In comparison to KV, however, majority of the available data on biological activities stems from studies using cells in vitro models, while studies based on observations in animals in vivo remains very limited.
Contrastingly, it appears that the vast majority of animal studies involved reasonable doses of amentoflavone (10-100 mg/kg) (Yu et al. 2017;Xiong et al. 2021). Additionally, most of the in vivo studies used positive control in comparable doses to amentoflavone (Chen et al. 2018;Cao et al. 2021). However, both the tested substance, as well as the positive control were in many cases administrated via different route than orally (e.g. intraperitoneally, subcutaneously) (Kim et al. 1998;Shin et al. 2006;Sakthivel and Guruvayoorappan 2013;Zhao et al. 2017Zhao et al. , 2019Chen et al. 2018;Liu et al. 2020;Rizk et al. 2021), indicating, that amentoflavone has a poor pharmacokinetics. On top of that, these routes do not correspond with the traditional application of G. kola. In terms of toxicity, amentoflavone is relatively well studied, and demonstrated inhibition towards several important enzymes of the human cytochrome P-450, including CYP 1A2, 2A6, 2B6, 2D6, 2C, 2E1 and 3A. The strongest inhibition was observed in the case of CYP2C8 and 2C9, where the IC 50 's were at 0.05 lg/mL (0.018 lM) and 0.08 lg/mL (0.15 lM), respectively. Other enzymes were inhibited in the range of 0.7-6.4 lg/ mL (1.3-11.9 lM) (Park et al. 2020). Additionally, amentoflavone was also found to inhibit numerous UDP-glucuronosyl transferases 0.06-9.08 lg/mL (range 0.12-16.86 lM) (Lv et al. 2018). Interaction with CYP-450 is associated with altered activity of some prescription drugs (e.g. St. John's Worth is known to interfere with such drugs as oral contraceptives, warfarin, digoxin, theophylline, indinavir, and cyclosporin) (Dewick 2009). Therefore, amentoflavone could be considered as an agent that potentially hinders activity of commonly prescribed drugs. Similar phenomenon have been observed for flavonoids associated with grapefruit (e.g. naringenin) (Fuhr et al. 1993). As far as we known, interference with CYP450 and UDP-glucuronosyl transferases was only observed for amentoflavone and remains unknown for other biflavonoids found in G. kola.
The research on volkensiflavone/morelloflavone shares many similarities with that of amentoflavone, except significantly lower number of studies on these compounds exist. However, it is again mainly restricted to in vitro studies. The available animal studies usually use only one dose (which in considerable share of studies is not based on application of pure compound but constitutes of mixture of structurally related compounds; for details see Table 5) and the administration route only involves intraperitoneal application. On top of that, as far as we know, positive control was used only in few studies. Even studies performed under in vitro seldomly uses it. In addition, from the clinical perspective, some of the studies present unrealistically high inhibitory concentrations (e.g. anti-cancer effect at levels 49.5 lg for morelloflavone and 100 lg/mL for amentoflavone) (Baggett et al. 2005), given that the in vitro efficiency of commonly used drugs (e.g. TaxolÒ) is in the range of ng/mL (nM) levels (Altmann and Gertsch 2007). Pharmacological efficiency of volkensiflavone/morelloflavone is thus questionable.
To the best of our knowledge, there are only two studies dealing with biological activity of garcinianin (Ajayi et al. 2014;Ito et al. 1999). It was tested for antibacterial effect against various pathogenic bacteria. Since garcinianin is found in roots and these are traditionally used as chewsticks for oral hygiene, this biological activity perfectly follows the ethnomedicinal indication. Garcinianin was found to be active against Streptococcus mutans, however, at enormously high dose (MIC = 1.0 mg/mL), being some 1000 times higher than commonly used antibiotics (Rubin et al. 2011). Additionally, it also showed potential anti-tumour promoting activity by inhibiting activity against 12-O-tetradecanoylphorbol-13-acetate (TPA)-induced Epstein-Barr virus early antigen activation in Raji cells at quite low concentrations (see Table 8).
Biological activity of G. kola biflavonoids is often being linked to the antioxidant-related mechanism (e.g. free radical scavenging activity, increased endogenous antioxidant defences, such as catalase, superoxide dismutase, and glutathione S-transferase). However, dietary antioxidants (including common flavonoids) have mostly failed to provide preventative and therapeutic activity in clinical studies of human disease, as reviewed in Halliwell (2012). The reasons behind flavonoids' inactivity may lie in their high hydrophobicity, and resulting low bioavailability. On top of that, these molecules are present in nearly every higher plant, including food plants. Humans are thus heavily exposed to these compounds, which may have resulted in development of efficient metabolization and elimination of these compounds from our bodies (Tauchen et al. 2020). There are few examples of dietary antioxidants which are believed to provide therapeutic benefit in oxidative stress-related diseases. One example of such compound is ergothioneine, which occurs in food sources in a relatively small quantities, yet human body accumulates it efficiently in various tissues. Ergothioneine is transported to the sites of accumulation via very specific transporter OCTN1. During illness (such as neurodegenerative, eye, and cardiovascular disorders) ergothioneine blood levels are significantly decreased suggesting that a deficiency could be relevant to the disease onset or progression. None of these features have been thus far observed for flavonoids (Halliwell et al. 2018;Cheah and Halliwell 2021). Apart from antioxidantrelated mode of action, G. kola biflavonoids have also been suggested to interfere with other systems, such as those involving inhibition of COX, phospholipase A2, aromatase, PDE, AChE, MAO-A, HMG-CoA reductase enzymes (see section biological activity of particular compound). For example, amentoflavone inhibits COX-1 and PDE at levels of 6.7 lg/mL (12.4 lM) (Bucar et al. 1998) and 0.15 lg/mL (0.27) (Saponara and Bosisio 1998), respectively. No inhibition was observed in the case of COX-2. However, in comparison to indomethacin 0.02 lg/mL (0.05 lM), amentoflavone shows some 250-fold COX-1 affinity (Kalgutkar et al. 2000) and to ViagraÒ approx. ten-fold PDE affinity (3.1 ng/mL; 6.6 nM) (Saenz De Tejada et al. 2001). Morelloflavone was shown to inhibit phospholipase A2 at IC 50 = 0.9 lM (Gil et al. 1997 (Lin et al. 2015). Since all G. kola biflavonoids are structurally related, similar affinities may be expected for all of them. The exact mechanism of action of these compounds (if there is any clinically relevant) still remains unknown. Of particular importance is also to note that many flavonoids have been marked as pan-assay interfering compounds (abbreviated as PAINs), providing false positive results in many enzymic assays, by virtue of their chemistry (e.g. inhibiting enzymes not by specific mechanism, but via production of radicals, such as H 2 O 2 ) (Bajorath 2021). Only very small number of flavonoid structures have successfully advanced to clinical use-examples of such compounds are intravenous silymarin in treatment of liver damage and injury (Ferenci 2016) and oral daflonÒ (mixture of micronized fraction of flavonoids, chiefly composed of 90% diosmin) (Lyseng-Williamson and Perry 2003). This indicates, that clinical efficiency of many flavonoids and their role in drug discovery remains questionable. Numerous in vitro and in vivo studies about biological activity of garcinol were conducted over the last few years. In comparison to some other compounds of G. kola, garcinol is usually used in reasonable doses. On the other hand, most of the studies lack the comparison with positive control. Moreover, in vivo studies often do not respect traditional way of administration. Its synergistic effect with conventional anticancer agents when administrated orally belongs to the most convincing results.
For example 0.05% garcinol in diet improved the response of transgenic pancreatic cancer mice to conventional treatment with gemcitabine from 10-15 to 25% (Saadat et al. 2018), Its combination with low dose of TaxolÒ was also able to better control the development of advanced or metastatic breast cancer (Tu et al. 2017). Additionally, it ameliorated the obesity-induced colon cancer, in this case, however, the i.p. application was better than the oral one. Garcinol also demonstrated in vitro antimicrobial activity against various G? bacteria on the same levels as conventionally used antibiotics. The effect against G-bacteria is significantly weaker and in case of E. coli it is even contradictory (MIC 25 lg/mL vs. 500 lg/mL) ( Table 6). Considering its significant activity against MRSA S. aureus strains, together with its potential to reduce the skin inflammation, the use of garcinol in topical treatments could become one of the research lead for this compound. A lot of evidence has been collected about garcinol positive effect against development and symptoms of Parkinson's disease (Deb et al. 2019). It reduced seizure scores, mortality rates and improved memory of PD mice in the same doses as valproate (Hao et al. 2016). Furthermore, its effect on dyskinesia of mice when administered orally was comparable to other natural HAT inhibitors as anacardic acid and curcumin which were administered i.p. and in up to 50 times higher doses (Ryu et al. 2018). Garcinol was shown to inhibit HAT at IC 50 of approx. 7 lM. (Balasubramanyam et al. 2004). The most potent reported HAT inhibitors identified so far are the bi-substrate inhibitors (e.g. H3-CoA-20; approx. IC 50 = 300 nM) (Lau et al. 2000). Another detail which may point to garcinol being possibly of value as a pharmaceutical agent is its striking structural resemblance to some already established drugs, such as hyperforin from St. John's Worth with antidepressant activity (Dewick 2009). However, despite some interesting research results about garcinol activity against PD and cancer, its pharmacokinetic properties have not been investigated in animal models yet. Therefore, its way to human clinical trials remains (at least for now) closed.
The research on garcinoic acid is very limited and chiefly constitutes of in vitro studies. Some of the presented effective doses, e.g. in the case of anticancer effect (Mazzini et al. 2009) are in the range (& 4.3 lg/mL; 10 lM) incomparable to conventionally used drugs (such as TaxolÒ, being efficient in nanomolar levels in in vitro tests) (Altmann and Gertsch 2007). The available animal studies use reasonable doses, though some are missing positive control (see Table 7). Even in majority of in vitro studies positive controls are not involved. As far as we know, only one study used oral administration (Marinelli et al. 2020). The remaining studies applied the compound via intraperitoneal or intradermal route, which is again not in correspondence with the traditional way of application. Number of studies dealing with garcinal is even more limited than in the case of garcionic acid and they are exclusively focused on determination of its antioxidant effect in vitro (Terashima et al. 1997(Terashima et al. , 2002. Since both garcinoic acid and garcinal are closely related to vitamin E, it has been suggested that they may provide therapeutic benefit through same antioxidant-related mechanism. However, it has been implied that mode of action of vitamin E may be derived from production of cell signaling and specific regulation of various genes rather than antioxidant activity (Azzi and Zingg 2005). This may also be true in the case of garcinoic acid and garcinal. Additionally, it has been previously shown that the above-mentioned biological effect is quite unique for a-tocopherol and the activity of related structures (e.g. b-, c-, and d-tocopherol) is significantly weaker (being 50%, 10%, and 3%, respectively) (Dewick 2009). This also applies for tocotrienols. It is therefore uncertain whether G. kola derived vitamin E derivatives have the capability to produce comparable pharmacological effect as a-tocopherol or their efficiency is significantly diminished as in the case of remaining a-tocopherol derivatives. Additionally, dtocotrienol, garcinoic acid and garcinal have displayed similar affinity towards PDE-5 as ViagraÒ, however, these results were only thus far observed in an in silico model (Ojo et al. 2021). Vitamin E have largely failed in clinical trials to provide therapeutic benefits in various human diseases (such as cardiovascular disorders, hypertension, diabetes, and cancer) (Robinson et al. 2006;Steinhubl 2008). Though generally recognized as safe, it has been found that high doses of vitamin E are associated with manifestation of various side effects, including hemorrhagic stroke (Sesso et al. 2008) and increased risk of prostatic cancer (Klein et al. 2011). In addition, vitamin E was suggested to interact with cytochrome P-450-dependent drug-metabolizing system (Brigelius-Flohé 2007), thus giving a one possible explanation to its ability to increase the blood thinning activity of warfarin (Fan et al. 2017). Quite recently, garcinoic acid was found to interfere with pregnane X receptor, which leads to regulation of cytochrome P-450 system . It thus appears, that apart from other adverse effects mentioned above, vitamin E and related structures (including garcinoic acid) may jeopardize therapeutic efficiency of some commonly used drugs.
Kolanone was mainly tested in in vitro antibacterial activity assays. As far as we known, only one study used isolated compound (Hussain et al. 1982); the remaining studies used extracts which under subsequent chemical analysis were found to contain kolanone. However, other compounds could also contribute to the observed antimicrobial effect. Again, there is a rationale behind testing kolanone for antibacterial effect, since it is also present in roots which are used as chewing sticks for oral hygiene. For the determination of antibacterial activity, Hussain et al. (1982) have used disc-diffusion methods and observed zones of inhibition at 14-15 mm for Pseudomonas aeruginosa, Staphylococcus aureus, Bacillus subtilis, Streptococcus pneumoniae, and Candida albicans. However, kolanone was applied in a solution of a very high concentration (1% w/v) without comparing its activity to proper positive control. Furthermore, the diffusion method is not appropriate for testing non-polar samples or those that do not easily diffuse into agar (Cos et al. 2006). Since kolanone appears to be derived from polyketide metabolism and is seemingly of non-polar nature, it may not be suitable for this kind of method. Recently, kolanone was tested in the rat model of ethanol-HClinduced gastric ulcers (Uwagie-Ero et al. 2020). It was administered in various doses (25, 50 and 75 mg/kg) and its efficiency was comparable to omeprazole (20 mg/kg). However, the mode of administration was not indicated. To the best of our knowledge, there is no study on the biological activity of gakolanone, a structurally related compound to kolanone. Since both kolanone and gakolanone are quite unusual constituents, and only limited share of studies were focused on them, questions may be raised about the veracity of deriving their structure. However, the structure was elucidated through UV, IR, MS, NMR techniques (Hussain et al. 1982;Akoro et al. 2020). Kolanone was also independently semi-synthetized in laboratory conditions (Raikar et al. 2008) (see Table 9).
For the remaining substances, garcifuran A and B, and garcipyran, the biological activity remains unknown. Only phytochemical records exist of these substances, and all of them have been thus far exclusively found only in G. kola (Niwa et al. 1994a, b). Again, as they represent a quite unique structure, and only very limited number of studies addressed them, concerns may be raised if their structures have been elucidated accurately. As in the case of kolanone and gakolanone, structures of both garcifurans and garcipyran have been elucidated via MS, IR, UV, and NMR. On top of that, there is one study reporting total synthesis of garcipyran B (Kelly et al. 1997). Pharmacological efficiency of these compounds may be questioned as well. However, some of the simple, as well as more complex, benzofuran derivatives are potentially of value as medicinal agents or are already in clinical use (such as griseofulvin, methoxalen, amiodarone, benziodarone, dimemebfe, efaroxan, elopiprazole).
Garcinia kola also contains phytosterols cycloartenol and 24-methylenecycloartenol, though they are quite abundant in the nature and are also found in other species including such genera as Artocarpus, Euphorbia, Costus, Polygonum, and Schinziophyton. In a recent study by Sadasivan Nair et al. (2020), both cycloartenol and 24-methylenecycloartenol showed glucose lowering activity in an oral glucose tolerance test in high fat diet-streptozotocin induced type II diabetic rats. Their antidiabetic activity might stem from ability to inhibit a-glucosidase (Nokhala et al. 2020). 24-Methylenecycloartenol was also found to attenuate acetic acid-induced pain in mice models of nociception (Ferreira et al. 2000). It also produced anti-inflammatory, antibacterial, and antiplasmodial effect (Akihisa et al. 1996;Bickii et al. 2006;Ajayi et al. 2014). Though showing some activities, both cycloartenol and 24-methylenecycloartanol are regarded as the starting structures and intermediates for the biosynthesis of other biologically active molecules (e.g. phytostanols, phytosterols) and are generally not considered as pharmacologically important compounds.
On top of so far discussed compounds, G. kola also contains exogenous constituents collectively referred to as cytochalasins, specifically 8-metoxycytochalasin J, cytochalasin H, cytochalasin J and alternariol, which  (1982) appears to be product of the endophytic fungi (Phomosis sp.) associated with the seed but are not synthesized by the plant itself. These compounds have been tested for antibacterial effect against various microorganisms, including Vibrio cholerae, Shigella flexneri, and Staphylococcus aureus and cytotoxic activity against HeLa cells, thought their efficiency was quite low (MIC ranged from 128 to 512 lg/mL and IC 50 against HeLa cells was in the range 0.25-35.69 lg/mL) (Jouda et al. 2016). Again, these activities are incomparably high to commonly used antibiotics and anticancer drugs (for references see above). Cytochalasin are quite recently discovered compounds. Their pharmacological value remains to be established.
As it is noted on several occasions in this review, KV is largely perceived as the active principle of G. kola. However, from what we know so far, other compounds present in G. kola, such as garcinol, garcifuran A and B, kolanone and gakolanone, may largely contribute to the bioactivities of G. kola and perhaps administer a greater promise for the drug discovery. However, none of the compounds found in G. kola have been subjected to the human clinical trials as of yet. Without them, any statement about what substance(s) is/are responsible for the biological activity of G. kola is a mere speculation. Although some compounds may display promising in vitro and in vivo activity, these results are unfortunately to a large extent not transferable to the clinical environment (as many compounds that were found to be effective in animal models later failed to provide sufficient action in humans) (Bracken 2009;Gertsch 2009), and this may also be the case of G. kola derived compounds.
There is a strong indication that G. kola possess some therapeutic benefits, as documented by its widespread use in folk medicine. Despite the numerous studies that have been conducted on bitter kola compounds, we still have little definite evidence of which substances are responsible for these therapeutic effects, nor do we know their exact mechanism of action. It is also quite possible that previously unknown substances are responsible for the biological activities of the species. There might be an analogy with turmeric (Curcuma longa), a traditional medicinal plant of Ayurveda, where curcumin has been identified as the active principle, yet available clinical studies have shown contradictory results. It appears that other, thus far unidentified compounds are responsible for the therapeutic benefit of the plant (Baker 2017).

Concluding remarks
Garcinia kola is an important medicinal plant with a long history of being used in the treatment of a wide range of human diseases. It contains several very specific compounds, which may be responsible for the observed biological activity and pharmacological properties of this plant. However, biological activity of these compounds, including perhaps the most studied substance kolaviron, has been only studied in animals. Confirmation that these substances are responsible for the therapeutic effects of the G. kola must be based on sufficiently powerful, double-blind, placebo-controlled clinical studies in humans (together with elucidation of their modes of action, therapeutic dose, adverse-effect profile, and other pharmacological data), which are unfortunately to date unavailable. We are afraid that at this moment therapeutic efficacy of any compound present in G. kola is far from conclusive. In connection to that, due to the relatively wide portfolio of diseases that are traditionally treated with G. kola and an even greater number of biological activities demonstrated by the present compounds, it is still impossible to reliably identify a substance that could be associated with the traditional ethnomedical use of G. kola. Many review articles have identified kolaviron as the active principle of G. kola. Perhaps garcinol, due to the relatively promising pharmacological activity (e.g. anticancer, antimicrobial, neuroprotective activities) deserves a deeper scientific interest. However, it is also likely that the substances potentially responsible for the pharmacological properties of the bitter kola have not yet been discovered. It is also possible that the constituents in G. kola work in synergy and, when isolated, will not provide such results as in the form of complex mixture in the natural material (as for example seen in the case of rauwolfia alkaloids). Hopefully some human clinical trials will be performed with the extracts/compounds from G. kola in the future and a promising candidate will emerge with the potential of becoming an important lead for the drug development.