Pharmacology, Toxicity, Bioavailability, and Formulation of Magnolol: An Update

Magnolol (MG) is one of the primary active components of Magnoliae officinalis cortex, which has been widely used in traditional Chinese and Japanese herbal medicine and possesses a wide range of pharmacological activities. In recent years, attention has been drawn to this component due to its potential as an anti-inflammatory and antitumor drug. To summarize the new biological and pharmacological data on MG, we screened the literature from January 2011 to October 2020. In this review, we provide an actualization of already known anti-inflammatory, cardiovascular protection, antiangiogenesis, antidiabetes, hypoglycemic, antioxidation, neuroprotection, gastrointestinal protection, and antibacterial activities of MG. Besides, results from studies on antitumor activity are presented. We also summarized the molecular mechanisms, toxicity, bioavailability, and formulations of MG. Therefore, we provide a valid cognition of MG.


INTRODUCTION
Magnoliae officinalis cortex, which was first recorded in "Shennong Herbal Classic" (Qin and Han Dynasty, around 221 B.C. to 220 A.D.), is the dry bark, root bark, and branch bark of Magnolia officinalis Rehd. et Wils. or Magnolia officinalis Rehd. et Wils. var. biloba Rehd. et Wils. In traditional medicine, Magnoliae officinalis cortex mainly acts to dry dampness and disperse phlegm, lower Qi, and eliminate fullness. Clinically, it is commonly used to treat asthma, constipation, edema, abdominal distension, malaria, and other diseases by combining different traditional Chinese medicines. For example, the Da Houpo Pill is used to treat abdominal distension (Song Ji Zonglu). The Xiaochengqi decoction is used for the treatment of tidal fever, constipation, and abdominal pain (Treatise on Febrile Diseases). The Banxia Houpo decoction has therapeutic effects on chronic pharyngitis, chronic bronchitis, and esophageal fistula (Synopsis of the Golden Chamber). Recent studies have shown that Magnoliae officinalis cortex has multiple pharmacological activities on the nervous system (Lee et al., 2009;Lee et al., 2013), digestive system (Kim HJ et al., 2018), inflammation (Kim JY et al., 2018), and cancer (Kim et al., 2020). And, its neolignan compounds include MG (a), honokiol (b), 4-methylhonokiol (c), and (R)-8,9dihydroxydihydromagnolol (d)   (Figure 1).
MG was first isolated from magnolia bark by Japanese scientist Sugii in 1930 and was first synthesized by Swedish scientist H. Erdtman and J. Runebeng with the p-allylphenol as raw material (Erdtman and Runeberg, 1957). However, the yield was only 25%, and it was challenging to separate and purify. Zhang et al. used 2,2′-biphenol and 1-bromobutane as raw materials to prepare MG (Zhang and Sun, 2011). The reaction process was simple and effective with mild conditions as well as high product purity (>98%), and the yield was increased to 60.2%.
The studies about MG's toxicity have been done, suggesting that MG has no genotoxicity and mutagenic toxicity (Saito et al., 2006). As a phenolic polyhydroxy compound, MG's poor aqueous solubility and low oral bioavailability limit its clinical use. Therefore, various formulations such as liposomes , solid dispersions (Stefanache et al., 2017a), emulsions (Sheng et al., 2014), and nanoparticles  have been developed to ameliorate the water solubility and bioavailability of it.
In this review, the pharmacological activities and molecular mechanisms of MG are summarized and updated. Its toxicities, bioavailability, and formulations are reviewed, to identify the benefit of further studies on MG and to find the best method to improve its bioavailability.

MATERIALS AND METHODS
This article collected literature studies related to pharmacology, toxicity, bioavailability, and formulation of MG published from January 2011 to October 2020. All related information about MG was collected by using the keyword of magnolol from globally recognized scientific search engines and databases, such as Web of Science, Springer, ScienceDirect, Elsevier, Google Scholar, and Chinese National Knowledge Infrastructure (CNKI). The source information of Magnoliae officinalis cortex was provided by the 2020 edition of Chinese Pharmacopoeia. The pharmacological activities, molecular mechanisms, toxicity, bioavailability, and formulations of MG are summarized, and the deficiencies of current studies are discussed. Inhibited iNOS and COX-2 expression and NF-κB activation via regulating PI3K/Akt and MAPK signaling pathways Lai et al. (2011) MTT-induced U937 cells In vitro: 10-100 μM Inhibited NO production and expression of p-IκBα, p-P65, IL-1β, and TNF-α. Downregulated phospho-JNK (p-JNK) and p-p38 Chen H et al.
The above results showed that MG has the effect of treating inflammation. However, most of the studies lacked positive groups. Positive groups should be set in follow-up studies.

Antitumor Activity
In the past few decades, in order to elucidate the molecular mechanisms of tumor formation and tumorigenesis and explore therapeutic methods, a mass of studies have been done. Currently, commonly used treatment methods include radiotherapy, chemotherapy, and surgery. However, present chemotherapeutic drugs have adverse reactions such as vomiting, hair loss, kidney damage, and bone marrow destruction. It is an important challenge to find effective and economic antitumor drugs with minimum side effects. A large number of literature studies have shown that MG has antitumor activity against colon cancer (Kang et al., 2012;Park et al., 2012), prostate cancer , liver cancer, lung cancer (Seo et al., 2011;Shen et al., 2017), gastric cancer (Rasul et al., 2012), cholangiocarcinoma (Zhang FH et al., 2017), oral cancer (Hsieh et al., 2018), ovarian cancer (Chuang et al., 2011), breast cancer , and melanoma (Cheng et al., 2020). MG suppressed the growth, migration, and invasion of tumor cells and promoted apoptosis as well as autophagy by acting on caspase-8, caspase-3, and other proteins participated in the p53, MAPK, NF-κB, TLR, HIF-1α/VEGF, PI3K/Akt/ERK/ mammalian target of rapamycin (mTOR), and Wnt/β-catenin signaling pathways Liu et al., 2013;Li et al., 2015;Shen et al., 2017;Zhang P et al., 2017).
In vivo, MG (5-20 mg kg −1 , i.p. injection) inhibited the growth of GBC-SD tumor in BALB/c nude xenograft model. It significantly increased caspase-3 activation and inhibited cell division cycle gene (CDC) 2 expression . In addition, treated with MG (40 mg kg −1 , i.p. injection) in the nude immune-deficient mice, it could be observed that the growth of nude immune-deficient MDA-MB-231 and MCF-7 tumors was inhibited, and the level of MMP-9 was decreased . In the human GBM orthotopic xenograft model, compared with temozolomide, cotreatment with MG and honokiol could more effectively inhibit tumor progression and induce apoptosis (Cheng et al., 2016).
In a word, MG and honokiol suppress the proliferation, migration, and invasion of tumor cells and promote apoptosis as well as autophagy by regulating MAPK, NF-κB, HIF-α, PI3K/ Akt/ERK/mTOR, and Wnt/β-catenin signaling pathways (Tse et al., 2005;Vavilala et al., 2014;Lin et al., 2016;Lee et al., 2019). In addition, MG shows antitumor activity by regulating TLR signaling pathways. Honokiol also can regulate STAF, EGFR, and notch signaling pathways to exhibit antitumor activities (Leeman-Neill et al., 2010;Liu et al., 2012;Kaushik et al., 2015). Further experiments in vivo are needed, and attention should be paid to whether MG could cause side effects.

Antiangiogenesis Activity
Angiogenesis, the essential procedure of embryonic angiogenesis, organ regeneration, and wound healing, is involved in many pathological illnesses, such as cancer, rheumatoid arthritis, and diabetic retinopathy. It is of great significance to study the molecular mechanism of angiogenesis, find relevant new drugs, and provide potential lead candidates. Studies have shown that ROS can participate in the signal transduction cascade in the key steps of angiogenesis and regulate the growth and migration of endothelial cells. MG inhibited angiogenesis through regulating the PI3K/Akt/mTOR signaling pathway and HIF-1α/vascular endothelial growth factor (VEGF)dependent pathway and inhibiting ROS production (Kim GD et al., 2013;Chen et al., 2013).
MG (10 μM) reduced the accumulation of HIF-1α protein by enhancing the activity of prolyl hydroxylase and reducing the synthesis of HIF-1α protein . MG (20 μΜ) has been shown to significantly inhibit the transcription and translation activity of platelet endothelial cell adhesion molecules and induce the production of ROS by mediating mitochondria and apoptosis. Furthermore, MG inhibited the activation of MAPKs and PI3K/Akt/mTOR signaling pathways in mouse embryonic stem (MES)/embryoid body (EB)-derived endothelial-like cells (Kim GD et al., 2013). MG (10 and 40 μM) suppressed the proliferation of human umbilical vein endothelial cells (HUVECs), ERK1/2 activity, gelatinase activity, and production of ROS and promoted HO-1 levels (Kuk et al., 2017).

Cardiovascular Protection
Cardiovascular disease is a large class of diseases, including coronary artery disease, hypertension, dyslipidemia, congenital heart disease, valve disease, and arrhythmia. With the improvement of people's living standards, the incidence of cardiovascular diseases is gradually increasing. MG showed activities of inhibiting the migration and hyperplasia of vascular smooth muscle cells (VSMCs), such as antiplatelet, antithrombotic, and antihypertensive via inhibiting MAPK family activation, Akt/ERK1/2/GSK3 β-catenin pathway, and angiotensin-converting enzyme (ACE)/angiotensin II (Ang II)/ Ang II type 1 receptor (AT-1R) cascade and upregulating PPARβ/γ and NO/guanosine 3′,5′-cyclic phosphate/PKG pathways (Shih and Chou, 2012;Karki et al., 2013b;Liang X et al., 2014;Wu et al., 2015;Chang et al., 2018). Under pathological conditions, the proliferation and migration of VSMCs to the intima can lead to vascular diseases such as atherosclerosis and restenosis after balloon angioplasty (Karki et al., 2012). MG (20 and 30 μM) inhibited VSMCs migration, β1-integrin expression, focal adhesion kinase (FAK) phosphorylation, RhoA and cell division cycle 42 (Cdc42) activation, and collagen-induced stress fiber formation (Karki et al., 2013b). MG (20 μM) suppressed VSMC proliferation and DNA synthesis by inhibiting the expression of cyclin D1/E and cyclin-dependent kinase 2 and 4, ROS production, and activation of renin-angiotensin system, MEK, and ERK1/2 (Karki et al., 2013a;Wu et al., 2015). Additionally, it (1-100 μM) could play the role of vasodilator and eliminate superoxide anion by relaxing right coronary arteries (separated from hearts of pigs) in a dosedependent manner and controlling the expression levels of iNOS and COX-2, with an IC 50 value of 5.78 μM . Further pharmacological research in this field was needed to reveal the mechanism by which MG inhibited homocysteineinduced endothelium-dependent vasodilation damage.

Hypoglycemic Activity
Diabetes is a metabolic disease characterized by hyperglycemia, which is caused by insufficient insulin excretion and impaired biological effects. Long-term hyperglycemia can contribute to chronic injury and dysfunction in numerous tissues, especially eyes, kidneys, and heart. Type 2 diabetes, formerly known as adult-onset diabetes, mostly occurs after 35-40 years of age and accounts for more than 90% of diabetic patients (Maddaloni et al., 2020). Numerous studies have reported that MG exhibits the hypoglycemic activity and protein tyrosine phosphatase 1B (PTP1B) inhibition by mediating AMPK/silent information regulator 1 (SIRT1)/PGC-1α, PPAR-γ, and protein kinase A (PKA) pathways, enhancing the activities of glyoxalase 1, PDX1, Ins2, and GPX genes, stimulating Akt phosphorylation, and inhibiting α-glucosidase Wang HY et al., 2013;Onoda et al., 2016;Pulvirenti et al., 2017;Suh et al., 2017;Parray et al., 2018).
Low-dose MG (0.01-1 μM) inhibited the death of RIN-m5F cells and the decrease of insulin secretion induced by methylglyoxal, thereby exerting hypoglycemic activity (Suh et al., 2017). It could upregulate the expression of Ins2 and PDX1, the levels of SIRT1 and PGC1α, AMPK phosphorylation, and glyoxalase 1 activity. Moreover, it attenuated the level of methylglyoxal-modified protein adducts and protected protein glycosylation (Alonso-Castro et al., 2011). In L6 myotubes, honokiol (3-30 μM) and MG (3-30 μM) stimulated glucose uptake in a dose-dependent manner and promoted the translocation of glucose transporter-4 to the cell surface as well as Akt phosphorylation. Their activity to stimulate glucose uptake could be blocked by the phosphatidylinositol 3kinase inhibitor, wortmannin . MG (20 μM) reduced metabolic disorders, oxidative stress, and fat formation by promoting the adipocyte differentiation and browning of 3T3-L1 C3H10T1/2 cells adipocyte-specific marker genes (uncoupling protein 1, CD137, Tbx1, etc.) and protein expression (Parray et al., 2018). It upregulated key fatty acid oxidation and lipid biomarkers (carnitine palmitoyltransferase 1C, acyl-CoA synthase long-chain family member 1, SIRT1, and perilipin) and activated AMPK, PPAR-γ, and PKA pathways. Honokiol and MG inhibited α-glucosidase with IC 50 values of 2.3 and 0.4 µM, respectively (Wang HY et al., 2013). Moreover, their inhibition at 1.5 μM was 3.9 and 29.8%, respectively . The inhibitory effect of honokiol on α-glucosidase was lower than that of MG.
C57BL/6J mice were fed a high-fat diet (45 kcal% fat) with or without honokiol (0.02%, w/w) or MG (0.02%, w/w) for 16 weeks. The results showed that honokiol and MG significantly lowered the weight of white adipose tissue, adipocyte size, and proinflammatory gene expression, protected against insulin resistance, and elevated plasma IL-10 level. In particular, honokiol could significantly decrease the plasma resistin level and increase the plasma adiponectin level compared to the control group (Kim YJ et al., 2013).
It can be seen that MG and honokiol have similar mechanisms to play a hypoglycemic role, such as inhibition of α-glucosidase and stimulation of glucose uptake. The difference is that MG has a better inhibitory effect on α-glucosidase, while honokiol can significantly decrease the plasma resistin level and increase the plasma adiponectin level.

Gastrointestinal Protection
In vitro, MG (3-100 μM) inhibited the spontaneous contraction, acetylcholine (ACh)-and Bay k8664-induced contraction, L-type Ca 2+ current, and the contraction of colonic smooth muscle through decreasing L-type Ca 2+ channel activity .
In the Kunming mouse model of diarrhea induced by castor oil, MG (25, 50, and 100 mg kg −1 , gavage) significantly inhibited diarrhea, reduced small intestinal transport, and increased catalase (CAT), superoxide dismutase (SOD), and glutathione peroxidase (GSH-Px) (Pang et al., 2013). Zeng et al. found that the antidiarrheal mechanism of MG and honokiol was similar, but in vivo experiments showed that MG had a higher antidiarrheal activity than honokiol (Zeng et al., 2015). The reason might be related to the inhibition of the liver CYP450 enzyme. Deng et al. reported that MG (100, 300, and 500 mg kg −1 , gavage) and honokiol (100, 300, and 500 mg kg −1 , gavage) regulated the release of IP3-Ca2+ storage, suppressed SK channel, and facilitated the opening of BKα1 as well as BKβ3 channels and the closing of BKβ4 channel by blocking the IP3-Ca2+ channel, inhibiting the activation of IP3 receptor 1 and CaM, and regulating protein kinase C (PKC) (Deng et al., 2015). In this study, the dose of MG and honokiol was too high and there was no positive control, so the dose should be reduced, and a positive control should be set for further research.
In conclusion, both MG and honokiol can exhibit gastrointestinal protective activity with similar mechanism, while MG's antidiarrheal activity is better than that of honokiol.
In BV2 cells, MG (10 μM) attenuated Aβ-induced AD by inhibiting the luciferase activity of NF-κB and the target gene of inflammatory cytokines, activating luciferase and liver X receptor activity, reducing ROS production induced by Aβ, upregulating apolipoprotein E (ApoE), and promoting microglial phagocytosis and Aβ degradation (Xie et al., 2020). MG (EC 50 3.49 μM) and honokiol (EC 50 2.65 μM) promoted the transcriptional activities of PPAR-γ in a dose-dependent manner. They also dose-dependently increased the luciferase activity of PPARγ-LBD. MG and honokiol could fit into the protein pocket of PPAR-γ-LBD with IC 50 values of 3.745 and 16.13 μM, respectively. What is more, MG had two hydrogen bonds at Glu343, which maintained the binding stability, while honokiol had one hydrogen bond at Glu343 and SER342, respectively, indicating that MG was more effective in enhancing PPAR-γ luciferase levels than honokiol (Xie et al., 2020). MG (5 μM) significantly inhibited trimethyltin (TMT)-mediated neuronal death and microglial activation by inhibiting ROS production and the activation of JNK, p38 MAPKs, and NF-κB in HT22 cells and BV-2 cells (Kim and Kim, 2016). Both MG (12.5 μM) and honokiol (6.25 μM) showed effective behavioral and electrophysiological antiepileptic activities in pentylenetetrazole and ethyl ketopentenoate models (Li G et al., 2020).
At concentrations of 50 and 100 mg kg −1 , MG alleviated depression-like behavior in male ICR mice by reducing corticosterone (CORT) level and increasing NE, 5hydroxytryptamine (5-HT), and BDNF protein levels (Bai et al., 2018). It could improve depressive behavior and hippocampal nerve damage in male ddY mice (Matsui et al., 2016). The phosphorylation of Akt, ERK, and cyclic AMPresponsive element-binding protein was significantly increased. In a male Kunming mouse model of chronic mild stress (CMS), MG (20 and 40 mg kg −1 , gavage) downregulated the levels of interleukin-1β (IL-1β), IL-6, and tumor necrosis factor-α (TNFα) in the prefrontal cortex, suppressed the activation of microglia and the proliferation of HPA axis and oxidative stress, and reversed malondialdehyde increase and SOD as well as GSHPx decrease to produce antidepressant-like effect (Cheng et al., 2018). MG (10 and 30 mg kg −1 , i.p. injection) downregulated the expression of bax and Ac-FOXO1 and production of NOS, 4-HNE, iNOS, phosphorylated p38MAPK, and C/EBP homologs, while upregulated the expressions of Bcl-2 and SIRT1. The regulation effect of MG on ischemic damage factors may be through inhibiting the production of ROS and upregulating p-Akt and NF-κB . MG (40 and 80 mg kg −1 ) exhibited antiepileptic activity by prolonging the latency of seizure onset and decreasing the number of seizure spikes, through acting on GABAA/benzodiazepine receptor (Chen CR et al., 2011). As indicated by the above results, both MG and honokiol can act on CB1 and CB2 receptors. The difference is that MG is a partial agonist of CB1 and CB2, while honokiol is a full agonist of CB1 and an inverse agonist of CB2, and MG has no activity on GPR-55, while honokiol is an antagonist of GPR-55. MG and honokiol can improve both phasic and tonic GABAergic neurotransmission in hippocampal dentate granule neurons; however, honokiol has a stronger positive regulatory effect on GABAA receptors than MG. In addition, MG and honokiol promote the transcriptional activities of PPAR-γ in a dosedependent manner. They also dose-dependently increased the luciferase activity of PPAR-γ-LBD. However, MG is more effective in enhancing PPAR-γ luciferase levels than honokiol. MG had antidepressant, anti-AD, anticonvulsant, antineurological deterioration, and protective effects to brain injury in the nervous system. Honokiol can regulate CB2 receptor, PPAR-γ targets, GABAA, and NF-κB and inhibit the levels of IL-Iβ, IL-6, IL-8, and TNF-α, production of ROS, RNS, COX2 as well as iNOS, and expression of PI3K/Akt, MAPKs, ERKs, JNKs, and p38 to exert neuroprotective effects (Talarek et al., 2017).

Interaction with CYP450 Enzyme
CYP450 is an important enzyme system involved in drug metabolism in vivo (Totah and Rettie, 2005). Among them, CYP2C8, CYP2C9, CYP2E1, and CYP2A6 accounted for about 40% of the total CYP450 enzymes in the liver (Zhang P et al., 2017). It is of great significance to study the interaction between the active components of traditional Chinese medicine and CYP450 for clinical safety. Studies have shown that MG can inhibit many CYP enzymes in humans and rats.

Antibacterial Activity
According to the literature review, MG has antibacterial activities. It could inhibit the Aeromonas hydrophila strains, with the minimal inhibitory concentration (MIC) value range of 32-64 μg ml −1 (Dong et al., 2017). MG and honokiol exhibited similar inhibitory activity against methicillin-resistant Staphylococcus aureus (MRSA) and methicillin-susceptible S. aureus (MSSA), with the MIC/minimal bactericidal concentration (MBC) value range of 16-64 mg L −1 (Zuo et al., 2015). Honokiol and MG dose-dependently inhibited the MRSA strain with the MIC values of 33 and 20 μg ml −1 , respectively (Kim SY et al., 2015). They inhibited multidrug-resistant and MRSA with MIC values in the range of 8-16 ppm . Choi et al. reported that honokiol and MG caused significant cellular immune-modulatory effect and decreased the production of ROS and inflammatory cytokines/ chemokines during S. aureus infection. Honokiol upregulated type I and III interferon mRNA expression in response to MSSA infection and inhibited the growth of MSSA at 2.5 μg ml −1 and MRSA at 5 μg ml −1 , whereas MG inhibited the growth of both bacterial cells at 5 μg ml −1 after 24 h of growing (Choi et al., 2015). MG and honokiol could inhibit S. mutans to prevent dental caries, with an MIC value of 10 μg ml −1 . And MG (50 μg ml −1 ) had better bactericidal activity against S. mutans biofilm than honokiol (50 μg ml −1 ) and chlorhexidine (500 μg ml −1 ) at 5 min after exposure (Sakaue et al., 2016).
Oufensou et al. tested the antifungal activities of MG and honokiol (5-400 μg ml −1 ) against 32 Fusarium spp. strains. The terbinafine (0.1-10 μg ml −1 ) and fluconazole (1-50 μg ml −1 ) were used as positive controls. The results revealed that MG had similar bactericidal activity compared with fluconazole, whereas honokiol had a better effect of inhibiting the mycelium growth compared to this fungicide. Compared to terbinafine, honokiol exhibited similar antifungal activity, whereas MG was less effective at all selected concentrations (Oufensou et al., 2019).

Antioxidant Activity
Amorati et al. explored the chemistry behind the antioxidant activity of MG and honokiol. They found that MG trapped four peroxyl radicals, with a kinh of 6.1 × 104 M −1 s −1 in chlorobenzene and 6.0 × 103 M −1 s −1 in acetonitrile, while honokiol trapped two peroxyl radicals in chlorobenzene (kinh 3.8 × 104 M −1 s −1 ) and four peroxyl radicals in acetonitrile (kinh 9.5 × 103 M −1 s −1 ). Their different behavior was due to the combination of intramolecular hydrogen bonding among the reactive OH groups (in MG) and of the OH groups with the aromatic and allyl π-systems (Amorati et al., 2015). MG has a bisphenol core with two allylic side chains, and its antioxidant activity is attributed to hydroxyl and allyl groups (Baschieri et al., 2017). MG downregulated myeloperoxidase (MPO) activity and the expression of TNF-α, iNOS, and IL-6 by altering JNK/ mitochondrial/caspase and PI3K/MEK/ERK/Akt/FoxO1 signaling pathways Dong et al., 2013).
It was found that MG (20 mg g −1 , i.v. injection) could significantly reduce MPO activity and the expression of iNOS, TNF-α, and IL-6 to inhibit oxidative stress and reduce mesenteric reperfusion caused lung injury in male C3H/HeOuJ mice . In aristolochic acid (AA)-induced HK-2 cells, MG (10 μM) and honokiol (10 μM) effectively reduced oxidative stress and suppressed cell proliferation by blocking the cell cycle at the G1 phase and preventing the G2/M arrest (Bunel et al., 2016).

Other Activities
Besides these pharmacological activities mentioned above, MG also has the following activities: inhibition of osteoclast differentiation, antiphotoaging, antiparasitic, antiviral activity, and reduction of multidrug resistance.
After treating HR-1 hairless male mice with 40 μL of the 0.25% MG preparation, it significantly reduced the average length and depth of wrinkles and inhibited the expression of MMP-1, MMP-9, and MMP-13 to play a role in antiphotoaging activity (Im et al., 2015).
MG significantly inhibited HBV activities. The IC 50 values of HBV surface antigen (HBsAg), HBV e antigen (HBeAg), and replication of HBV DNA were 2.03, 3.76, and 8.67 μM, respectively, and without cytotoxicity to HBsAg and HBeAg . MG (2.51 ± 0.51 μg ml −1 ) and honokiol (3.18 ± 0.61 μg ml −1 ) stimulated the expression of immunerelated genes to resist grass carp reovirus infection in Ctenopharyngodon idella kidney (CIK) cells. MG significantly increased the expression of interferon (IFN) regulatory factor (IRF) 7 and IL-1β to activate type I IFN (IFN-I) but failed to induce the molecules in NF-κB pathways. The difference was that honokiol promoted the expression of IL-1β, TNFα, NF-κB, IFN-β, promoter stimulator 1, IRF3, and IRF7 but failed to increase IFN-I expression, showing that it could enhance the host innate antiviral response to grass carp reovirus infection by regulating NF-κB pathway .
What is more, MG (1-50 μM) reduced the multidrug resistance of cancer cells to antitumor drugs through downregulating P-glycoprotein expression in a concentrationand time-dependent manner and increased the intracellular accumulation of calcein in NCI/ADR-RES cells (Han and Van Anh, 2012).

TOXICITY
So far, a large number of studies have shown that MG has cytotoxicity ( Table 2). MG (10-100 μM, 24 or 48 h) was used to investigate the toxicity on human normal hepatocyte U937 and LO-2 cells. The results showed that MG at low concentration could promote the cell survival rate in a dose-dependent manner. At a concentration of less than 60 μM, MG could promote the survival of U937 cells. When exposed to MG at a concentration of less than 70 μM after 48 h, the mortality of LO-2 cells was lower than 20% . Additionally, at a concentration range from 50 to 200 μg ml −1 , MG could cause toxicity and inhibit MMEC survival (Wei et al., 2014). Karki et al. reported that MG at a concentration of 40 μM possessed cytotoxicity on VSMCs (Karki et al., 2013a). MG (100 μM) reduced the murine 3T3-F442A preadipocyte viability by 25% and human normal subcutaneous preadipocyte viability by 36%. MG (50 μM) reduced the murine cell viability by 16% and human cell viability by 22%. Otherwise, honokiol (50 μM) significantly decreased the murine and human cell viability by 30 and 39%, and the combined application of honokiol and MG (100 μM each) markedly decreased the cell viability by 73% (murine) and 80% (human). The combined application of honokiol and MG (50 μM each) also markedly reduced murine (31%) and human (37%) cell viability. On the contrary, the simultaneous application of honokiol and MG (30 μM each) only moderately affected the murine (15%) and human (21%) cell viability (Alonso-Castro et al., 2011). When the concentration of MG was > 50 μM, it would be toxic to mES-derived endothelial-like cells (Kim GD et al., 2013) 52 ± 5.09, 36.46 ± 2.38, 59.40 ± 8.24, 35.69 ± 4.91, 25.39 ± 3.26, 25.32 ± 2.72, and 24.79   . When OC2 cells were treated with MG (20-100 μM) for 24 h, the cell viability decreased in a dose-dependent manner (Hsieh et al., 2018). After treating A549 cells with 6. 25, 12.5, 25, 50, 100, and 200 μM of MG for 24 and 48 h, cell viability for 24 h was 98.1 ± 2.7, 86.4 ± 2.3, 79.5 ± 4.6, 68.7 ± 2.3, 55.9 ± 1.1, and 12.8 ± 3.1%, respectively, while for 48 h was 92.5 ± 3.5, 80.1 ± 4.7, 70.2 ± 2.8, 56.6 ± 3.4, 36.3 ± 2.6, and 3.1 ± 0.9%, respectively. When the dose of MG was ≤6.25 μM, there was almost no inhibitory effect on A549 cells, while 25 μM of MG significantly inhibited the proliferation of A549 cells. MG inhibited the proliferation of A549 cells in a dose-and timedependent manner . In DU145 cells, the viability was reduced by 30 and 60% at 40 and 80 μM, respectively, after 6 h of MG treatment, and 49 and 76% were reduced at 40 and 80 μM, respectively, after 24 h of MG treatment. After treating PC3 cells with 80 μM MG for 6 and 24 h, its viability decreased to 50 and 48%, respectively . Li et al. treated GBC cells with MG at concentrations of 10, 20, and 30 μM for 48 h. The results showed that the apoptosis index of GBC cells was significantly higher than that of the control group . SGC-7901 cells were treated with different concentrations of MG (0, 10, 30, 50, 100, 200, and 300 µM) for 48 h. It was observed that MG inhibited cell growth in a dose-dependent manner. Compared with the control group, exposing cells to 40, 60, and 80 µM of MG for 48 h resulted in a significant reduction in the number of cells (Rasul et al., 2012). MG significantly suppressed the proliferation of SKOV3 and TOV21G cells in a dose-dependent (6.25, 12.5, 25, 50, and 100 μM) and time-dependent (48 and 72 h) manner (Chuang et al., 2011). The QBC939, SK-ChA-1, MZ-ChA-1, and RBE cells were treated with different concentrations of MG (20, 40, 80, and 160 μM) for 24, 48, and 72 h. The results demonstrated that MG significantly suppressed the proliferation of the above cell lines in a concentration-and time-dependent manner (Zhang FH et al., 2017).
Fujita et al. investigated the ability of MG and honokiol to inhibit UV-induced mutation in Salmonella typhimurium TAI02. The results suggested that both MG (5 μg/per plate) and honokiol (5 μg/per plate) could inhibit against UV-induced mutations by scavenging ·OH generated by UV irradiation. The relative mutagenic activities of MG and honokiol were 62 ± 1% and 62 ± 4%, respectively, while that of control was 100% (Fujita and Taira, 1994). MG significantly inhibited the mutagenicity induced by indirect mutagens but did not affect the direct mutagens. It strongly and competitively inhibited the activities of ethoxyresorcinol-O-demethylase and methoxyresorcinol-Odemethylase, indicating that it could inhibit indirect mutageninduced mutations by suppressing the activities of CYP1A1 and CYP1A2 (Saito et al., 2006). The genotoxicity of Magnolia bark extract (MBE) was studied by Li et al., which was composed of 94% MG and 1.5% honokiol. The results revealed that MBE was not genotoxic under the conditions of the in vitro bacterial reverse mutation test and in vivo micronucleus test and supported the safety of MBE for dietary consumption (Li et al., 2007).
In general, the abovementioned cytotoxicity is mostly related to the antitumor and antiangiogenic activities of MG. Additionally, studies have shown that MG not only has no mutagenic and genotoxic activity but also even has antimutagenic activity. In summary, MG was found to be fairly nontoxic.

BIOAVAILABILITY AND FORMULATION
MG is a dimeric phenolic neolignan  with strong lipid solubility, and its absorption in the gastrointestinal tract is mainly through a lipid-like pathway (Niu et al., 2015). Hatorri et al. studied the absorption, metabolism, and excretion of MG through oral administration and intraperitoneal injection of [ring-14 C] MG. The results showed that MG participated in enterohepatic circulation (Hattori et al., 1986). After oral administration of MG (50 mg kg −1 ), the MG sulfates and glucuronides were predominant in the bloodstream. And MG was mainly distributed in the liver, kidney, brain, lung, and heart; among these organs, the concentration of MG and MG glucuronides in the liver was the highest . Additionally, MG's main metabolite excreted in bile was magnolol-2-O-glucuronide, and the main route of excretion of MG after oral or intraperitoneal injection was through the alimentary tract (Hattori et al., 1986). After 24 h of oral administration of [ring-14 C] MG, the main fecal derivatives of oral MG in rats were MG and a series of free form metabolites, which accounted for more than 90% of the total dose; only 6% were glucuronic acid and sulfate (Hattori et al., 1986). The MG metabolites tetrahydromagnolol and trans-isomagnolol showed an increasing trend after repeated administration, indicating that their formation was related to the induction of metabolic enzymes in animal tissues and/or intestinal bacteria. It was mainly excreted through liver metabolism and renal excretion (Hattori et al., 1986). The absorption half-life, elimination half-life (T 1/2 ), maximum concentration-time (T max ), and maximum concentration (C max ) of MG were 0.63 h, 2.33 h, 1.12 h, and 0.16 μg ml −1 , respectively. The water solubility and gastrointestinal absorption of MG were poor, with the oral bioavailability of only 4.9% (Tsai et al., 1996), limiting its clinical use. The low bioavailability might be partly due to the high metabolism of the intestine and liver and the low solubility in gastric juice.
Liu et al. prepared MG solid dispersion, MG solid lipid nanoparticles, and MG phospholipid complex and studied their bioavailability. The results showed that the cumulative dissolution of MG was 30.6% within 12 h, while the cumulative dissolution of MG solid dispersion, MG solid lipid nanoparticles, and MG phospholipid complex increased to 96.3, 76.4, and 45.9%, respectively. The pharmacokinetic parameters such as C max and area under the curve (AUC) 0-t and AUC 0-∞ were significantly improved. Moreover, compared with raw MG, their relative bioavailability increased to 1.38, 2.12, and 3.45 times, respectively (Liu et al., 2020). All three preparations could improve the oral absorption bioavailability of MG, but the effect of MG solid lipid nanoparticles was more obvious. Lin et al. prepared a solid dispersion of MG with polyvinylpyrrolidone K-30 (PVP) and studied its bioavailability by oral administration (50 mg kg −1 ). The results indicated that compared with raw MG, the solid dispersion of MG with PVP significantly increased the systemic exposures of MG and MG sulfates/glucuronides by 80.1 and 142.8%, respectively (Lin et al., 2014). For the solid dispersion prepared by MG and croscarmellose sodium (1: 5), the in vitro cumulative dissolution rate of MG reached 80.66% at 120 min, which was 6.9 times that of the raw MG (11.74%) (Tang et al., 2016). Stefanache et al. incorporated MG into the pores of amino-functionalized mesoporous silica particles to increase the dosage of MG and delay its release (Stefanache et al., 2017a).
After gavaging the emulsion (50 mg kg −1 ) in male SD rats, the 1.20 h average plasma concentration of MG was 426.4727 ng ml −1 , and the absolute bioavailability was 17.579%, indicating that preparing an emulsion could improve the bioavailability of MG (Sheng et al., 2014).
Mixed Soluplus (SOL) and Solutol HS15 (HS15), SOL, and D-alpha-tocopheryl polyethylene glycol 1,000 succinate (TPGS) were used to prepare MG-loaded mixed micelles (MG-M) MG-H and MG-T, respectively. The relative oral bioavailability of MG-T and MG-H were increased by 2.39-and 2.98-fold, respectively, compared to that of raw MG, indicating that MG-H and MG-T could promote the absorption of MG in the gastrointestinal tract . Shen et al. also prepared MG-M by pluronic F127 and L61, and its drug loading efficiency and entrapment efficiency were 81.57 ± 1.49% and 27.58 ± 0.53%, respectively. In vitro release test showed that MG had sustained release behavior after being encapsulated in micelles. The permeability of MG through the Caco-2 cell monolayer was enhanced, and the relative bioavailability of oral MG-M was 2.83 times higher than that of the raw MG (Shen H et al., 2018). It can be seen that the mixed micelle drug delivery system can improve the poor water solubility and bioavailability of MG.
In general, the existing formulations can not only improve the water solubility and bioavailability of MG but also improve its stability, enhance its pharmacological effects, and enable MG to have a sustained release behavior, which will provide strategies for future clinical applications of MG.

CONCLUSION
In 2011, Chen et al. summarized the pharmacological activities and molecular mechanisms of MG. According to the review, MG could exhibit anti-inflammatory activity by inhibiting the production of inflammatory enzymes/cytokines and activation of NF-κB and leukocyte. It also exerted antitumor effects by inhibiting cell proliferation and metastasis and inducing apoptosis. The molecular mechanisms mainly include the increase of p21, p27, caspase-3, caspase-8, and caspase-9 expression, inhibition of PI3K/ PTEN/AKT pathway, ERK1/2, NF-κB, P38, iNOS, and COX2 activation, CYP1A1, CYP1A2, MMP-9 as well as MMP-2 activity and Bcl-2 expression, induction of cytochrome C, and AIF release and activation of the mitochondrial death receptor pathway. MG could attenuate VCAM-1, ICAM-1, MCP-1, and MMP9, inhibit the proliferation of smooth muscle cells and fibroblasts, and obtain arrhythmia from I/R injury to show cardiovascular protection. It could also exert neuroprotective activities by inhibiting the production of PGE2, regulating (GABA) A receptor subtypes and central serotonergic activity, retaining cholinergic neurons in the forebrain, and inhibiting cortical 5-HT release. MG had a therapeutic effect on gastrointestinal diseases by regulating serotonergic and gastrointestinal system functions and relaxing gastrointestinal smooth muscles. Moreover, it exhibited hypoglycemic activity by activating PPAR and increasing basal and insulin-stimulated glucose uptake (Chen YH et al., 2011).
In this review, in vivo and in vitro studies demonstrated that MG has a wide range of pharmacological activities including anti-inflammatory, antitumor, antioxidant, hypoglycemic, cardiovascular protection, antiangiogenesis, and antibacterial. MG inhibited TLR2/TLR4/NF-κB/MAPK/PPAR-γ pathways and decreased the expression of inflammatory cytokines to exhibit anti-inflammatory activity. It suppressed the growth, migration, and invasion of tumor cells and promoted apoptosis as well as autophagy, through acting on caspase-8, caspase-3, and other proteins participated in the p53, MAPK, NF-κB, TLR, PI3K/Akt/mTOR, and Wnt/β-catenin signaling pathways. It also protected the nervous system through multiple systems and multiple targets. Moreover, it has a wide range of antibacterial activity. MG is a candidate drug for anti-inflammatory, anticancer, and neuroprotective activities. However, MG's in vivo effect with CYP enzymes is not clear yet, and there is no clinical research on MG, which cannot fully provide the pharmacological activities of it.
MG and honokiol have similar pharmacological activities. Both of them can exhibit antitumor activities by regulating MAPK, NF-κB, HIF-α, PI3K/Akt/ERK/mTOR, and Wnt/β-catenin signaling pathways. MG shows antitumor activity by regulating TLR signaling pathways, and honokiol can regulate STAF, EGFR, and notch signaling pathways to exhibit antitumor activities. They have inhibitory activity on α-glucosidase and stimulation of glucose uptake to play a hypoglycemic role, while MG has a better inhibitory effect of α-glucosidase. Moreover, both MG and honokiol exhibit gastrointestinal protective activity with similar mechanism, while MG's antidiarrheal activity is better than that of honokiol. MG is a partial agonist of CB1 and CB2; however, honokiol is a full agonist of CB1 and an inverse agonist of CB2. MG has no activity on GPR-55, while honokiol is an antagonist of GPR-55. Honokiol has a stronger positive regulatory effect on GABAA receptors than MG; however, MG is more effective in enhancing PPAR-γ luciferase levels than honokiol. What is more, the inhibition types of MG on CYP1A, CYP2C19, CYP2C, CYP3A, and CYP1A2 were competitive inhibition. The inhibition type of honokiol on CYP1A2 was competitive inhibition, and the inhibition types of honokiol on CYP2E1 and CYP2C19 were noncompetitive inhibition. Both honokiol and MG have antimicrobial activity. The difference is that honokiol exhibits better antimicrobial activity than MG on Aggregatibacter actinomycetemcomitans, S. mutans, S. aureus, MRSA, Escherichia coli, and Fusarium spp.
MG is nontoxic and is used in dietary supplements and cosmetic products, such as added to toothpaste to play antibacterial and antiperiodontitis effects. However, the low water solubility, poor bioavailability, and skin irritation hamper its application. To overcome this problem, numerous studies have been conducted. By preparing solid dispersions, nanoparticles, phospholipid complexes, liposomes, emulsions, etc., the bioavailability and stability of MG significantly improved, which will greatly promote its clinical application. Aside from its formulations, structural modification is becoming an increasingly promising method for obtaining MG derivatives with better therapeutic effects and higher bioavailability. The synthesis and research of MG derivatives are beyond the scope of this study, so we will not go into details. Consequently, the design and research of MG derivatives are of great significance in the future.
In summary, this article comprehensively reviews the pharmacology, toxicity, bioavailability, and formulations of MG.

AUTHOR CONTRIBUTIONS
YL and YS contributed to the conception and design of the study; YL, YL, and YZ prepared the original draft; BT, XQ, and QY reviewed and edited the manuscript; YS supervised the study.